- Email: [email protected]
SiC composite tubes under a prototypic high heat flux
Journal of Nuclear Materials 491 (2017) 94e104
Contents lists available at ScienceDirect
Journal of Nuclear Materials journal homepage: www.elsevier.com/locate/jnucmat
Experimental design and analysis for irradiation of SiC/SiC composite tubes under a prototypic high heat flux* Christian M. Petrie a, *, Takaaki Koyanagi a, Joel L. McDuffee a, Christian P. Deck b, Yutai Katoh a, Kurt A. Terrani a a b
Oak Ridge National Laboratory, P.O. Box 2008, Oak Ridge, TN, 37831, USA General Atomics, P.O. Box 85608, San Diego, CA, 92121, USA
a r t i c l e i n f o
a b s t r a c t
Article history: Received 30 August 2016 Received in revised form 11 April 2017 Accepted 20 April 2017 Available online 4 May 2017
The purpose of this work is to design an irradiation vehicle for testing silicon carbide (SiC) fiberreinforced SiC matrix composite cladding materials under conditions representative of a light water reactor in order to validate thermo-mechanical models of stress states in these materials due to irradiation swelling and differential thermal expansion. The design allows for a constant tube outer surface temperature in the range of 300e350 C under a representative high heat flux (~0.66 MW/m2) during one cycle of irradiation in an un-instrumented “rabbit” capsule in the High Flux Isotope Reactor. An engineered aluminum foil was developed to absorb the expansion of the cladding tubes, due to irradiation swelling, without changing the thermal resistance of the gap between the cladding and irradiation capsule. Finite-element analyses of the capsule were performed, and the models used to calculate thermal contact resistance were validated by out-of-pile testing and post-irradiation examination of the foils and passive SiC thermometry. Six irradiated cladding tubes (both monoliths and composites) were irradiated and subsequently disassembled in a hot cell. The calculated temperatures of passive SiC thermometry inside the capsules showed good agreement with temperatures measured post-irradiation, with two calculated temperatures falling within 10 C of experimental measurements. The success of this design could lead to new opportunities for irradiation applications with materials that suffer from irradiation swelling, creep, or other dimensional changes that can affect the specimen temperature during irradiation. Published by Elsevier B.V.
Keywords: Silicon Carbide Cladding Irradiation Design Composite
1. Introduction Silicon carbide (SiC) is an attractive material for use in nuclear applications due to its radiation tolerance and its strength and chemical inertness at high temperatures [1e4]. In addition to the inherent advantages of SiC, continuous SiC fiber-reinforced, SiC-
* This manuscript has been authored by UT-Battelle, LLC under Contract No. DEAC05-00OR22725 with the U.S. Department of Energy. The United States Government retains and the publisher, by accepting the article for publication, acknowledges that the United States Government retains a non-exclusive, paid-up, irrevocable, worldwide license to publish or reproduce the published form of this manuscript, or allow others to do so, for United States Government purposes. The Department of Energy will provide public access to these results of federally sponsored research in accordance with the DOE Public Access Plan (http://energy. gov/downloads/doe-public-access-plan). * Corresponding author. E-mail address: [email protected] (C.M. Petrie).
http://dx.doi.org/10.1016/j.jnucmat.2017.04.058 0022-3115/Published by Elsevier B.V.
matrix composites (SiC/SiC composites) offer increased damage tolerance, improved mechanical properties, and enhanced reliability compared to monolithic SiC [5e7]. These properties of SiC/ SiC composites make them intriguing candidates for use as structural materials for nuclear applications. One of the specific applications for using SiC/SiC composite materials is their utilization as accident tolerant fuel cladding and/or structural materials in light water reactors (LWRs) [8e13]. Other applications include components in advanced high-temperature gas-cooled reactors (HTGRs) [14], fluoride-salt-cooled high-temperature reactors (FHRs) [15], gas-cooled fast reactors (GFRs) [16,17], and fusion energy applications [2,18,19]. A number of architectures are being investigated for accidenttolerant fuel cladding applications, including fully composite or multilayered architectures incorporating a SiC/SiC composite material and a monolith [6,20]. An important concern is whether the SiC-based cladding is able to maintain hermeticity (and thus fission
C.M. Petrie et al. / Journal of Nuclear Materials 491 (2017) 94e104
product retention). Essentially, the structure is required to withstand the evolving stress state over the fuel lifetime when exposed to intense neutron irradiation under a prototypic high heat flux (and thus a large temperature gradient) as the rod internal pressure increases due to released fission gas buildup. Detailed discussions of the various governing phenomena and the nature of stress evolution in SiC-based cladding structures have been provided elsewhere [21e23]. Briefly, the high heat flux across the thickness of the cladding in LWR applications (~0.5e1 MW/m2) results in a large temperature gradient in SiC/SiC cladding due to the poor thermal conductivity of this composite that further degrades with radiation damage [6]. For metallic cladding, higher thermal conductivity reduces the radial temperature gradient and the associated stresses due to differential thermal expansion across the thickness. Therefore, the primary driver of stress (in the absence of pellet-cladding contact) is the pressure difference between the inside and outside of the cladding. However, in SiC/SiC cladding, the higher radial temperature gradient results in a much more important stress driver that dominates the strain profile across the cladding thickness: the differential swelling strain across the thickness of the cladding. This is the case because swelling is a strong function of temperature in the LWR cladding temperature regime [24]. A previous irradiation experiment investigating fuel-clad interaction with multilayered SiC cladding materials showed cracking near the ends of the cladding where the heat flux (and thus temperature) was lower [25]. Because swelling of SiC is greater at lower temperatures [26], the cladding ends underwent larger volumetric expansion. It is thought that irradiation swelling caused additional stresses as the cladding expanded into the stainless steel capsule housing. These stresses are clearly not truly representative of typical LWR conditions where a somewhat uniform coolant temperature of >300 C is present on the cladding outer surface. This work describes an improved irradiation capsule design aimed specifically at addressing the performance of SiC/SiC cladding in LWR conditions. The design allows for a SiC cladding tube surface temperature in the range of 300e350 C under a representative high heat flux (~0.6 MW/m2) during one cycle of irradiation in an un-instrumented “rabbit” capsule in the High Flux Isotope Reactor (HFIR). Both SiC/SiC composites and chemical vapor deposited (CVD) SiC monoliths have been irradiated, and additional architectures incorporating various combinations of composite and monolith are being considered. To simplify the design and postirradiation examination, the fuel is replaced with a surrogate molybdenum “heater” cylinder that is designed to generate enough heat from gamma heating in the HFIR to achieve the desired heat flux. Design calculations show that the tube surface temperature remains constant to within 12 C despite volumetric irradiation swelling of the tube by as much as 3%. The results of these experiments and future irradiation experiments utilizing the design described in this work will provide experimental validation of fuel performance codes and ultimately improve the fundamental understanding of evolving stress states in SiC/SiC composite materials due to differential thermal expansion and irradiation swelling. Also, specimens irradiated in this configuration are essential for understanding the possibility of microcrack formation in these materials as a result of stress and whether this microcracking will result in release of gaseous fission products. 2. Materials The assembled capsules were designed for insertion into the flux trap of the HFIR (see Section 3.1.1) for irradiation. Neutron and gamma radiation in the flux trap of the HFIR, roughly two orders of magnitude higher than a typical LWR core, cause heating of the
95
capsule materials. Thermal design calculations are performed to estimate capsule component temperatures during irradiation. Temperatures for this capsule are controlled by the geometry, fill gas, and surface roughness of the contacting components. The overall design and assembly of the rabbit capsules is summarized in Fig. 1. Rabbits refer to small sealed aluminum capsules that serve as the irradiation containment vehicle. Fig. 2 shows a picture of the cross-section of an assembled capsule. Molybdenum heater rods (i.e. dense gamma-absorbing cylinders) are inserted inside the cladding tubes. Titanium centering thimbles keep both the cladding and the heaters centered within the capsule housing. A tight-fitting aluminum sleeve is wrapped around the cladding, and an embossed aluminum foil fills the space between the sleeve and the capsule housing. The embossed aluminum foil is the key to this design as it can absorb the irradiation swelling of the cladding without causing significant stress to the cladding. These components make up one sub-assembly. Fig. 1 shows four sub-assemblies: two sub-assemblies that include 16 mm length cladding tubes, one sub-assembly that includes a 12 mm length cladding tube, and one sub-assembly that includes a 4 mm length cladding tube. Alternatively, as was done for the irradiated capsules described in this work, the capsule could be loaded with three sub-assembliesdeach with one 16 mm length cladding tube. A multiple subassembly design allows for multiple cladding tube specimens to be irradiated at once. Grafoil insulating discs (brown components in Fig. 1) are positioned at the top and bottom of the capsule interior to reduce axial heat losses. Table 1 summarizes the capsule components, materials, and nominal dimensions that are relevant to the thermal design. The SiC/SiC tubes were fabricated using a chemical vapor infiltration process at General Atomics. Hi-Nicalon Type S SiC fiber was used for the reinforcement. Two composite tubes with IDs GA-TGI-1 and GA-TGI-4 were investigated. Four monolithic high-purity CVD SiC (Dow Chemical Co.) tubes machined from a plate (IDs CVD-T1, CVDT2, CVD-T6, and CVD-T7) were also investigated. All specimens (both monolith and composite) had inner and outer diameter concentricity within 0.02 mm. Surface roughness was measured to be 0.3 mm root mean square (RMS) for all monolith specimens, 13 mm RMS for GA-TGI-1, and 1.7 mm RMS for GA-TGI-4. The thermometry (or passive temperature monitors) listed in Table 1 are rectangular pieces of CVD SiC positioned inside the molybdenum heaters between the two centering thimbles of a sub-assembly. The thermometry can be examined post-irradiation as a form of passive temperature monitoring to determine the temperature during irradiation [27]. 3. Experimental 3.1. Experimental facilities 3.1.1. HFIR irradiation facility The HFIR is a beryllium-reflected, pressurized, light-watercooled and moderated flux-trap-type reactor located at the Oak Ridge National Laboratory (ORNL). The core consists of aluminumclad involute-fuel plates, which currently utilize highly enriched 235 U fuel at a power level of 85 MW [28]. A typical cycle is 24e25 days. The reactor core, illustrated in Fig. 3, consists of two concentric annular regions, each approximately 61 cm in height. The flux trap is ~12.7 cm in diameter, and the outer fueled region is ~43.5 cm in diameter. The fuel region is surrounded by a beryllium annular reflector, which is in turn backed up by a water reflector of effectively infinite thickness. The reactor core assembly is contained inside a pressure vessel, which is located in a pool of water. The rabbits described in this work are placed in the flux trap of the HFIR in an outer ring target rod rabbit holder (TRRH) position
96
C.M. Petrie et al. / Journal of Nuclear Materials 491 (2017) 94e104
Fig. 1. Exploded computer-aided design view of the SiC cladding tube rabbit assembly.
Fig. 2. Cross-section of an assembled rabbit capsule. For scale, the SiC cladding tube outer diameter is 8.50 mm.
closest to the inner fuel elements. In a TRRH position, up to seven rabbits can be stacked in a target. These rabbits were placed at the reactor midplane (position TRRH-4). The values of the fast (neutron energy > 0.1 MeV) and total neutron flux at this position are 1.1 1015 n/cm2$s and 4.3 1015 n/cm2$s, respectively [30]. Running the SPECTER-ANL code (Argonne National Laboratory) [31] with an average displacements per atom (dpa) cross section of
260 b gives 2.3 dpa in SiC for a 24 day HFIR cycle. 3.1.2. Thermal contact conductance measurement system In order to validate the thermal design simulations (described later in Sections 4 and 5) used to calculate component temperatures during irradiation, experimental measurements of thermal contact conductance through an embossed foil were performed
C.M. Petrie et al. / Journal of Nuclear Materials 491 (2017) 94e104 Table 1 Capsule parts, materials, and nominal dimensions. Part
Material
Dimension
Nominal Value
Housing
Aluminum 6061
Inner diameter Outer diameter
9.52 mm 10.96 mm
Housing end cap Heater
Aluminum 4047 Molybdenum
Cladding
Silicon carbide
Sleeve
Aluminum 1100
Foil
Aluminum 1100
Inner diameter Outer diameter Inner diameter Outer diameter Inner diameter Outer diameter Thickness Depth Spacing
2.50 7.09 7.12 8.50 8.50 9.00 0.05 0.26 0.83
Centering thimble Insulator discs Thermometry
Ti-6Al4V Grafoil Silicon carbide
mm mm mm mm mm mm mm mm mm
97
first pumping down the vessel to ~930 Pa (7 torr) and then backfilling with either helium or neon fill gas. The rabbit capsule design described in this paper utilizes helium fill gas, but both helium and neon gases were tested in the validation experiments in order to test the contact conductance equations used for the capsule design using two different fill gases. While running chilled water through the cooling block, the heater power was ramped up in steps. By measuring the input power and the temperature difference of the surfaces that contact the foil, the contact conductance through the foil, h (W/m2-K), can be calculated as shown in Eq. (1): 00
h¼
q ; DT
(1)
00
Length Cross-section
12.0 mm 1.0 mm 0.5 mm
inside of a sealed, instrumented vessel. Fig. 4 shows a 3D model of the experiment with labeled components. The experiment is placed inside a sealed vessel with additional insulation above and around the perimeter of the experiment to prevent natural convection loops from forming inside the vessel. A cartridge heater is used to provide the heat input. The power to the heater is controlled using a variac and measured using a watt meter. The heat is conducted from the square protrusion of the heating block, through the embossed foil, to the cooling block, which is cooled with chilled (~5 C) water. Thermocouples are inserted into small holes in both the heating block and the cooling block near the surfaces that contact the foil. Additional thermocouples are located in the surrounding insulation to give some indication of heat losses via conduction to the vessel walls. The thermocouples located in the insulation generally read less than 30 C throughout the experiments, indicating that heat losses through the vessel walls to the ambient air are likely minor. 3.2. Thermal contact conductance measurement procedure Thermal contact conductance experiments were performed by
where q (W/m2) and DT are the heat flux through the foil and the temperature difference across the foil, respectively. In this case, DT is the temperature difference of the thermocouples located near 00 the two surfaces in contact with the foil, and q is simply the heater power divided by the cross-sectional area (494 mm2) of the square protrusion of the heating block. 4. Calculations for in reactor performance The thermal design calculations described in this paper determine capsule component temperatures resulting from reactor heating. The ultimate heat sink is the reactor primary coolant, which flows through the reactor core and cools the outer surface of the capsule housing. 4.1. Materials and material properties Material properties and thermal boundary conditions for this work are taken from independently reviewed design and analysis calculations that are maintained internally within ORNL. Material property data are primarily obtained from CINDAS [32], MatWeb [33], and various literature sources. In general, material properties recommended by Snead et al. [24] for CVD SiC were used for all SiC tubes included in the simulations. This includes temperature- and dose-dependent irradiation swelling, temperature-dependent
Fig. 3. Cross-section through the HFIR illustrating the primary experimental sites [29].
98
C.M. Petrie et al. / Journal of Nuclear Materials 491 (2017) 94e104
Fig. 4. Computer-aided design model of experiment for validation of foil contact conductance.
coefficients of thermal expansion, and temperature- and swellingdependent values for elastic modulus and thermal conductivity. Although the elastic modulus for the composite tubes will be different from that of the monoliths, this should not affect the thermal analyses. Thermal conductivity of the irradiated composite tubes was assumed to be 3.0 W/m-K based on previous irradiations of SiC/SiC composites [6]. The nominal swelling of the composite tubes was assumed to be 2.2% (0.73% linear swelling). The effects of varying the irradiation swelling on the thermal contact conductance from the cladding tubes to the capsule housing are discussed later in Section 5.1. No anisotropy was considered in these analyses.
4.2. Heat generation and thermal boundary conditions Peak heat generation rates, parameters for determining axial profiles for the heat generation rates, and convection parameters are summarized in Table 2. The fill gas for this design is helium. The local heat generation rate is estimated with the following profile:
2 z ; qðmaterial; zÞ ¼ qpeak ðmaterialÞ,exp
(2)
s
Where
Table 2 Thermal boundary conditions. Parameter
Value
Convective heat transfer coefficient Bulk coolant temperature Peak heat generation rate for Al-6061 Peak heat generation rate for Ti-6Al4V Peak heat generation rate for SiC Peak heat generation rate for molybdenum Peak heat generation rate for grafoil Correlating parameter (s)
47.1 kW m2 K1 52 C 32.5 W/g 35.2 W/g 32.9 W/g 43.3 W/g 33.7 W/g 30.07 cm
q ¼ local heat generation rate as a function of material and axial location, qpeak ¼ heat generation rate at the HFIR core midplane as a function of material, z ¼ axial location in HFIR, where the core midplane is at zero, and s ¼ correlating parameter.
4.3. Finite-element model Thermal calculations of capsule component temperatures were performed using the ANSYS (ANSYS, Inc.) finite-element software code package. Additional user-written macros were implemented within ANSYS to calculate thermal conductance between contacting parts. These macros account for solid spot conductance, gas gap conductance, and thermal jump distances. Details regarding the solution methods can be found in a previous publication [34]. Thermal contact conductance is extremely important in this particular capsule design, as the high heat flux results in large jumps in temperature at each contact interface. The contact conductance depends on, among other parameters, the thermal conductivity of the fill gas and the surface roughness values for the parts that are in direct contact (i.e., the cladding tube, the sleeve, the foil, and the capsule housing). See Ref. [34] and the references therein (including [35e37]), for details regarding the thermal contact conductance models. The effects of contact pressure on the solid spot conductance and the gas gap conductance are neglected in this analysis as the contact pressures due to clad swelling and foil compression are likely negligible because the foil is very thin (0.05 mm), is very ductile since it is made of a fully annealed 1100 aluminum alloy, and is expected to experience significant irradiation creep. The effects of changing the effective contact area for heat transfer between the foil and the contacting components as the cladding swells and the foil crushes were explicitly modeled. It is later shown in section 5.1 that swelling of the cladding and crushing of the foil reduced the cladding outer surface temperature
C.M. Petrie et al. / Journal of Nuclear Materials 491 (2017) 94e104
by 12 C or less. In order to model the swelling of the cladding and the compression of the embossed aluminum foil, a 2D transverse slice is taken through one sub-assembly inside the rabbit. The 2D model also assumes 10 azimuthal symmetry. A full 3D modeling of foil compression was not attempted due to the small mesh size required to simulate foil compression as the cladding swells. The 2D model is used to determine the effective thermal contact conductance between the cladding and the capsule housing. Then, a 3D model is generated with the foil and sleeve suppressed. The gas gap between the clad and the housing is artificially modified until the contact conductance through this effective gas gap is equal to the calculated contact conductance in the 2D model (see Section 5.1) that includes the sleeve and the compressed aluminum foil. This design is heavily dependent on contact resistance at the clad/sleeve interface, the sleeve/foil interface, and the foil/housing interface. This is why the contact conductance through the embossed foil is validated experimentally. The contact conductance depends significantly on the surface roughness of the contacting surfaces. The surface roughness values of the inner surface of a standard rabbit housing and of an aluminum foil were measured to be 0.26 and 0.65 mm RMS, respectively, using a profilometer. The standard rolls of aluminum foil that are used to fabricate the sleeve have a surface roughness similar to the housing inner surface and the foil (i.e., ~0.3e0.7 mm RMS). The surface roughness of the cladding tube specimens can vary from as low as ~0.3 mm RMS for the CVD monolith specimens to as high as ~13 mm RMS for some composite specimens, as mentioned in Section 2. The surface roughness of the aluminum sleeve was varied in such a way that the effective contact conductance is approximately constant despite variations in clad surface roughness, thus maintaining the desired clad outer surface temperature. All tube specimens are paired with an 8.4 mm RMS sleeve except for composite specimen GA-TGI-1 (13 mm RMS roughness), which uses a 0.3 mm RMS sleeve.
99
The results obtained from the 2D simulations are used to calculate the effective contact conductance from the clad, through the sleeve, through the foil, and to the housing after clad swelling. This is calculated by dividing the heat flux at the outer surface of the clad (0.66 MW/m2) by the change in temperature from the clad outer surface (339 C) to the housing inner surface (68 C). With this method, the contact conductance is calculated to be 2.42 kW/ m2-K. This contact conductance is used in the 3D finite-element simulations (results shown in Section 5.3), which consider the effects of axial heat loss and spatially dependent heat generation. Note that because of the variations in the surface roughness of the cladding specimens and the sleeves, there is some variation in the effective contact conductance from the cladding to the housing. The value of 2.42 kW/m2-K was calculated for the GA-TGI-4 composite tube with an 8.4 mm RMS sleeve. Similar simulations, performed with the GA-TGI-1 composite specimen and with the monolith specimens (and their corresponding sleeves) resulted in contact conductance values of 2.32 and 2.49 kW/m2-K, respectively. To determine the sensitivity of the design temperatures to clad swelling, the linear clad swelling is varied from 0 to 1.05% (from 0 to 3.15% volumetric swelling) for the case with composite tube GATGI-4 and an 8.4 mm RMS sleeve. The 3.15% maximum volumetric swelling that is assumed is the maximum swelling of CVD SiC in the point-defect swelling saturation regime (150e1000 C) [24]. The cladding tube specimens tested in this work were later observed (see Section 6.1) to have linear swelling of up to 0.73% after one 25 day cycle of irradiation in the HFIR. The area-averaged outer surface temperature of the cladding is plotted vs. linear clad swelling in Fig. 6. If the linear swelling was to vary between 0 and 1.05%, the cladding outer surface temperature would be constant to within 12 C. In this scenario, the embossed aluminum foil design is effectively utilized to render the magnitude of the cladding swelling inconsequential to its temperature distribution. 5.2. Experimental validation of thermal contact conductance models
5. Results 5.1. 2D simulation results The 2D simulations are performed to determine the effective contact conductance from the cladding tubes to the capsule housing. Fig. 5 shows temperature contour plots obtained from 2D simulations with composite tube GA-TGI-4 (1.7 mm RMS and an 8.4 mm RMS sleeve). A linear clad swelling of 0.73% (~2.2% volumetric swelling) was assumed for the results shown in Fig. 5. Table 3 summarizes the temperatures of the various components. The heat flux at the outer surface of the cladding was calculated to be 0.655 MW/m2 in the 2D simulations.
Sections 3.1.2 and 3.2 describe the experimental setup and
Table 3 Part temperatures obtained from 2D simulations of composite tube GATGI-4 with 0.73% linear clad swelling. Part
Temperature ( C) Average (MineMax)
Centering thimble Heater Cladding Sleeve Foil Housing
608 (608e609) 601 (596e604) 415 (339e496) 304 (303e305) 135 (131e138) 66 (65e68)
Fig. 5. Contour temperature plot ( C) obtained from 2D simulations of composite tube GA-TGI-4 with 0.73% linear clad swelling.
100
C.M. Petrie et al. / Journal of Nuclear Materials 491 (2017) 94e104
Fig. 6. Composite clad tube GA-TGI-4 outer surface temperature vs. clad swelling determined from 2D simulations.
procedure for the out-of-pile experiments that were performed to validate the thermal contact conductance models used in the thermal design of the rabbit capsule. Figs. 7 and 8 show temperatures near the surface of the heating block (TC1) and the cooling block (TC2), respectively, as well as the heater power as a function of time during experiments with helium (Fig. 7) and neon (Fig. 8) fill gases. It is clear that the temperatures had essentially reached a steady-state for each of the three steps in heater power with both fill gases. Note that the brief decreases in temperature in Fig. 7 during the first 5e10 min were due to cooling of the system with chilled water prior to powering the heater. Using the steady-state temperatures for each step in power in Figs. 7 and 8, a plot of heat flux vs. (TC1 e TC2) or DT can be generated. Performing a linear fit to this data gives contact conductance through the foil for each fill gas. A simple 2D finiteelement model of the experimental rig was constructed to enable a direct comparison of the thermal contact conductance models with the experimental results. Note that the same contact conductance models are used in the experimental validation model and in the rabbit capsule design model. The surface roughness of the heating and cooling blocks was measured to be ~0.95 mm RMS, while the foil roughness was measured to be ~0.63 mm RMS. This information was input in the contact conductance equations in the finite element models. Fig. 9 shows heat flux vs. temperature difference (between the heating and cooling blocks), as determined from the validation
Fig. 7. Temperatures near the surface of the heating block (TC1) and the cooling block (TC2), as well as the heater power, as a function of time during the validation experiments with helium fill gas.
Fig. 8. Temperatures near the surface of the heating block (TC1) and the cooling block (TC2), as well as the heater power, as a function of time during the validation experiments with neon fill gas.
experiments and the finite element models using both helium and neon fill gases. The models show very good agreement with the experimental results. Using error propagation of uncertainties in the thermocouple measurements, power measurements, and dimensional tolerances, uncertainties in the measured values for contact conductance were determined to be ~1e3%. Note that the contact conductance in the validation experiment (~3.9 kW m2 K1 for helium fill gas) differs significantly compared to the contact conductance through the foil in the rabbit irradiation experiment (~2.4 kW m2 K1). This discrepancy is due to differences in surface roughness of the contacting components and the fact that the conductance for the rabbit irradiation capsule includes an extra contact resistance from the cladding/sleeve interface.
5.3. 3D simulation results With the contact conductance models validated via experimental measurements, 3D thermal models of the rabbit capsule were used to determine temperature contours for all of the capsule components. As mentioned in Section 5.1, the 3D models do not include the embossed foil or the sleeve. Instead, an effective gas gap
Fig. 9. Heat flux vs. temperature difference (between the heating and cooling block) data determined from the model and from validation experiments with helium and neon fill gases. Slopes indicate contact conductance through the foil.
C.M. Petrie et al. / Journal of Nuclear Materials 491 (2017) 94e104
between the cladding and the housing was varied manually until the contact conductance from the cladding to the housing was equal to the value determined by the 2D simulations; these 2D simulations do include the foil and sleeve and also account for clad swelling. The effective helium gas gap required to give a contact conductance of ~2.4 kW/m2-K is 83 mm. Fig. 10 shows temperature contour plots obtained from 3D models of the rabbit capsule with composite tube GA-TGI-4 and an 8.4 mm RMS roughness sleeve. The temperature at the outer surface of the tube specimen ranges from 296 to 322 C, with an average temperature of 317 C. Note that the region with relatively low temperature at the top of the tube specimen is due to the fact that the molybdenum heater was designed to be slightly shorter than the tube specimen so that it would not expand more than the tube specimen, which would risk pushing up on the centering thimbles. Therefore, the top of the tube specimen experiences a much lower heat flux. The temperature monitor inside the molybdenum heater has an average temperature of 580 C. This temperature can be compared to the results of post-irradiation dilatometry analyses that determine the actual temperature of the temperature monitor during irradiation. Similar 3D simulations were performed for composite tube GA-TGI-1 and for the CVD monolith tubes. Table 4 summarizes temperature distributions in the tube specimens and temperature monitors for all cases. All simulations resulted in an average clad surface temperature between 310 and 325 C. Temperature gradients through the cladding tubes varied significantly due to differences in the thermal conductivity of the irradiated tubes, which affected the temperatures of the thermometry placed inside the capsule sub-assemblies. The calculated thermometry temperature for the various cladding tubes is discussed in Section 6.2. 6. Discussion The experimental design described in this paper offers a new approach to controlling temperature in uninstrumented rabbit irradiation capsules. The typical approach is to rely on a small (on the order of tens to hundreds of microns, depending on the design) insulating gas gap that requires tight machining tolerances and is sensitive to things such as concentricity misalignment and irradiation swelling. The approach described in this paper relies on surface roughness and the contact area of the embossed foil. It is less sensitive to concentricity misalignment due to the self-
101
centering nature of the foil, and it does not require extremely tight machining tolerances on the outer diameter of the cladding because the foil is designed to crush when the cladding is inserted. This approach is particularly attractive for pressurized creep experiments or other applications where swelling, creep, or other phenomena can affect the dimensional stability of capsule components that comprise an insulating gas gap. 6.1. Post-irradiation examination Initial information regarding the performance of the irradiation capsules utilizing the embossed foil design has been acquired, and a more detailed post-irradiation examination of the tubes will be discussed in a future paper. Of the six specimens that were loaded into the two irradiated capsules (four CVD monoliths and two SiC/ SiC composites), five specimens survived the irradiation and postirradiation cutting operations. The authors believe that the single broken specimen (CVD-T6) was damaged during the postirradiation milling operations required to retrieve the specimens from the capsule housing. Specimen CVD-T2 broke later during post-irradiation handling; it may or may not have also been damaged during post-irradiation disassembly. Fig. 11 shows a picture of one of the rabbit housing cut open by milling, a post-irradiation picture of a monolith tube specimen, and a post-irradiation picture of a composite tube specimen. The periodic lines that appear on the inner surface of the housing indicate that the raised surfaces of the foil remained in contact with the housing as intended. The tube specimens show no signs of surface cracking or other degradation on the outer surface. The shiny sections on the outside of the tubes indicate locations where the aluminum sleeve stuck to the tubes, perhaps during the milling operation to split the capsule housing. Note that specimen GA-TGI1 generally had a very rough surface but it also had a section that was much smoother. These surface variations may have had an effect on the temperatures, which will be discussed in Section 6.2. Dimensional measurements of the cladding tubes were taken before and after irradiation to determine the total diametrical and axial swelling. Table 5 summarizes the irradiation swelling of the tube specimens. Note that specimens CVD-T2 and CVD-T6 were broken, and therefore the post-irradiation length could not be measured. Swelling of the length and outer diameter are mostly consistent, and all of the CVD tubes generally show similar linear swelling (~0.4e0.5%). Composite tube GA-TGI-4 shows similar
Fig. 10. Temperature ( C) contours showing composite clad tube GA-TGI-4 and heater sub-assembly (left), clad only (center), and clad outer surface (right) obtained from 3D simulations.
102
C.M. Petrie et al. / Journal of Nuclear Materials 491 (2017) 94e104 Table 4 Calculated temperatures for each tube specimen and the passive temperature monitors located inside the specimen sub-assemblies. Temperature ( C) Average (MineMax)
Specimen ID
CVD-T1 CVD-T2 CVD-T6 CVD-T7 GA-TGI-1 GA-TGI-4
Specimen
Specimen Outer Surface
Temperature Monitor
334 347 347 334 394 385
310 323 323 310 325 317
470 489 489 470 587 580
(298e367) (308e379) (308e379) (298e367) (304e479) (296e471)
(298e315) (308e325) (308e325) (298e315) (305e329) (297e321)
(469e471) (487e490) (487e490) (469e471) (583e589) (575e581)
Fig. 11. Picture of rabbit housing inner surface after hot cell disassembly (left) and post-irradiation pictures of a monolith tube (center) and composite tube (right).
swelling to the CVD specimens, whereas composite tube GA-TGI-1 is significantly higher (~0.7%). A KEYENCE VHX-1000 optical microscope was used to photograph the embossed foils and determine the profile of the foils before and after irradiation. The disassembly process damaged the foils in certain locations, so information could only be gathered in locations where the tools used in the hot cells did not damage the foil. 3D images were constructed by combining the single images with a different working distance. The step of the working distance was 3 and 10 mm for unirradiated and irradiated specimens, respectively. 2D profiles were obtained by taking an appropriate slice through the 3D image. Fig. 12 shows the 2D profile of a flat embossed foil pre-irradiation and a curved foil post-irradiation. The depth of the foil is approximately 238 ± 4 mm pre-irradiation, and 207 ± 17 mm post-irradiation. The difference in these values (~30 mm) represents the amount that the foil was compressed. This number corresponds well with the observed swelling of the tube specimens (18e30 mm).
6.2. Comparison of predicted vs. measured capsule temperatures Passive SiC temperature monitors were placed inside the molybdenum heaters to give a comparison of the actual irradiation temperature vs. the temperatures predicted from the design calculations. By analyzing the instantaneous coefficient of thermal expansion (CTE) of the temperature monitor during postirradiation isochronal heating and cooling, the median, minimum, and maximum temperatures during irradiation can be determined [27]. Fig. 13 shows the temperatures predicted from the passive thermometers that were placed inside each of the six irradiated tube sub-assemblies compared with the temperatures predicted from the thermal calculations. The experimental values shown are the median temperatures determined from analysis of the postirradiation isochronal annealing data. In general, the predicted temperatures show good agreement with the experimental values. The experimentally determined temperatures inside the four CVD
Table 5 Measured tube specimen irradiation swelling, accurate to within ±0.06%. Material type
Specimen Linear swelling of ID length
Linear swelling of outer diameter
Unirradiated length (mm)
Irradiated length (mm)
Unirradiated outer diameter (mm)
Irradiated outer diameter (mm)
CVD tube CVD tube CVD tube CVD tube SiC/SiC tube SiC/SiC tube
CVD-T1 CVD-T2 CVD-T6 CVD-T7 GA-TGI-1
0.39% N/A N/A 0.51% 0.73%
0.47% 0.55% 0.41% 0.48% 0.70%
16.03 16.03 16.03 16.03 15.95
16.093 N/A N/A 16.112 16.067
8.52 8.52 8.52 8.52 8.57
8.560 8.567 8.555 8.561 8.630
GA-TGI-4 0.46%
0.52%
16.02
16.094
8.54
8.585
C.M. Petrie et al. / Journal of Nuclear Materials 491 (2017) 94e104
103
7. Conclusions
Fig. 12. Measured 2D profiles of a flat unirradiated foil and a curved irradiated foil.
Fig. 13. Temperatures of passive thermometry included inside tube specimen subassemblies, determined from post-irradiation dilatometry, compared with temperatures predicted from thermal calculations.
monolith tubes ranged from 472 to 540 C, while the two composite tube temperatures were 502 and 586 C. The calculated temperature (470 C) for CVD-T1 matches the experimental value to within 2 C. The calculated temperatures for the three remaining CVD tubes are lower than the experimental values by as much as 57 C. For the composite tubes, the calculated temperature (580 C) for GA-TGI-4 matches the experimental value to within 6 C; on the other hand, the predicted temperature (587 C) for GA-TGI-1 is 85 C higher than the experimental value. Given the high heat flux in this design, the temperature at the location of the passive thermometry is sensitive to many variables, including the cladding tube thermal conductivity, tube swelling, and thermal contact conductance between the tubes and the capsule housing. With these considerations, the discrepancies (85 C for the composite tubes and up to 57 C for the monolith tubes) between calculation and experiment are not surprising. Thermal conductivity of the irradiated composite tubes is particularly uncertain given that reported values range from 1.2 to 3.9 W/m-K when the irradiation temperature and dose were in the range of 300e500 C and 0.8e71 dpa, respectively [6]. A more likely explanation is that the relatively smooth section of specimen GA-TGI-1 (see Fig. 11) resulted in increased thermal contact conductance in that section of the specimen, which would result in reduced temperatures. Ongoing work for future publication includes experimental investigations of the irradiation temperature within the SiC tubes based on the dimensional recovery behavior during annealing.
This paper describes a vehicle for testing silicon carbide tube materials under conditions representative of a LWR (constant clad surface temperature of 300e350 C under a high heat flux of ~0.6 MW/m2) in order to validate thermo-mechanical models of stress states in these materials due to differential thermal expansion and irradiation swelling. The design can accommodate clad swelling up to ~3% while keeping the clad surface temperature constant to within 12 C throughout the irradiation. An engineered aluminum foil was developed to absorb the expansion of the cladding due to irradiation swelling. This new approach to controlling temperature (as opposed to the traditional approach utilizing a small insulating gas gap) in uninstrumented rabbit irradiation capsules relies on surface roughness and the contact area of the embossed foil. It is less sensitive to concentricity misalignment due to the self-centering nature of the foil, and it does not require tight machining tolerances on the outer diameter of the cladding because the foil is designed to crush when the cladding is inserted. For the application described in this work, the high heat flux requires the tube specimen to have a relatively constant outer surface roughness in order to prevent variations in temperature due to non-uniform heat transfer from regions of varying surface roughness. This approach is particularly attractive for pressurized creep experiments or other applications where swelling, creep, or other phenomena can affect the dimensional stability of capsule components that would otherwise define an insulating gas gap. A finite-element model of the capsule performance was developed using ANSYS with custom libraries used for calculating thermal contact resistance. The thermal models were validated by out-of-pile testing. Initial post-irradiation examination of the cladding tube specimens reveals that five of the six irradiated tubes survived the irradiation and hot cell disassembly. Inspection of the capsule housing's interior surface and the embossed foils suggests that the design performed as expected. The calculated temperatures of passive SiC temperature monitors placed inside the capsules generally showed good agreement with the temperatures measured post-irradiation via dilatometric analysis, with two calculated temperatures agreeing with experimental values to within less than 10 C. Acknowledgements Lance Snead and Koroush Shirvan at MIT, Greg Markham at Ceramic Tubular Products, and Edgar Lara-Curzio at ORNL provided significant input and constructive criticisms on the irradiation vehicle design. Without the persistent support and encouragement of Frank Goldner at the United States Department of Energy (U.S. DOE) to proceed with this complex design, the completion of this irradiation would not have been possible. Padhraic Mulligan (ORNL) and Kevin Field (ORNL) provided useful comments on the manuscript. Doug Stringfield (ORNL) performed most of the capsule assembly. This research was supported by the Advanced Fuels Campaign of the U.S. DOE, Office of Nuclear Energy. Neutron irradiation in the HFIR was made possible by the Office of Basic Energy Sciences, U.S. DOE. References [1] L.L. Snead, T. Nozawa, M. Ferraris, Y. Katoh, R. Shinavski, M. Sawan, Silicon carbide composites as fusion power reactor structural materials, J. Nucl. Mater. 417 (1e3) (2011) 330e339. [2] L. Giancarli, H. Golfier, S. Nishio, R. Raffray, C. Wong, R. Yamada, Progress in blanket designs using SiCf/SiC composites, Fusion Eng. Des. 61e62 (2002) 307e318.
104
C.M. Petrie et al. / Journal of Nuclear Materials 491 (2017) 94e104
[3] Y. Katoh, L.L. Snead, I. Szlufarska, W.J. Weber, Radiation effects in SiC for nuclear structural applications, Curr. Opin. Solid State Mater. Sci. 16 (3) (2012) 143e152. [4] R. Konings, Comprehensive Nuclear Materials, Elsevier Science, 2011. [5] Y. Katoh, L.L. Snead, C.H. Henager Jr., T. Nozawa, T. Hinoki, A. Ivekovi c, S. Novak, S.M. Gonzalez de Vicente, Current status and recent research achievements in SiC/SiC composites, J. Nucl. Mater. 455 (1e3) (2014) 387e397. [6] Y. Katoh, K. Ozawa, C. Shih, T. Nozawa, R.J. Shinavski, A. Hasegawa, L.L. Snead, Continuous SiC fiber, CVI SiC matrix composites for nuclear applications: properties and irradiation effects, J. Nucl. Mater. 448 (1e3) (2014) 448e476. [7] T. Koyanagi, K. Ozawa, T. Hinoki, K. Shimoda, Y. Katoh, Effects of neutron irradiation on mechanical properties of silicon carbide composites fabricated by nano-infiltration and transient eutectic-phase process, J. Nucl. Mater. 448 (1e3) (2014) 478e486. [8] K. Yueh, K.A. Terrani, Silicon carbide composite for light water reactor fuel assembly applications, J. Nucl. Mater. 448 (1e3) (2014) 380e388. [9] D.M. Carpenter, An Assessment of Silicon Carbide as a Cladding Material for Light Water Reactors, Nuclear Science and Engineering, Massachusetts Institute of Technology, Cambridge, MA, 2011. [10] K.A. Terrani, B.A. Pint, C.M. Parish, C.M. Silva, L.L. Snead, Y. Katoh, Silicon carbide oxidation in steam up to 2 MPa, J. Am. Ceram. Soc. 97 (8) (2014) 2331e2352. [11] Y. Lee, T.J. McKrell, C. Yue, M.S. Kazimi, Safety assessment of SiC cladding oxidation under loss-of-coolant accident conditions in light water reactors, Nucl. Technol. 183 (2) (2013) 210e227. [12] Y. Katoh, K.A. Terrani, Systematic Technology Evaluation Program for SiC/SiC Composite-based Accident-tolerant LWR Fuel Cladding and Core Structures: Revision 2015, Oak Ridge National Laboratory, Oak Ridge, TN, 2015. [13] K.E. Barrett, M.P. Teague, I.J. Van Rooyen, S.M. Bragg-Sitton, K.D. Ellis, C.R. Glass, G.A. Roth, K.M. McHugh, J.E. Garnier, G.W. Griffith, Engineering challenges of LWR advanced fuel cladding technology in preparation for inreactor demonstrations, Trans. Am. Nucl. Soc. (2012) 886e889. [14] L.L. Snead, Y. Katoh, W. Windes, K. Smit, Candidate structural materials for incore VHTR application, Trans. Am. Nucl. Soc. (2008) 1019e1020. [15] C. Forsberg, L. Hu, Design goals for a fluoride-salt-cooled high-temperature test reactor (FHTR), Trans. Am. Nucl. Soc. (2013) 599e602. [16] A. Kohyama, T. Hinoki, T. Mizuno, T. Kunugi, M. Sato, Y. Katoh, J.S. Park, R&D of advanced material systems for reactor core component of gas coolded fast reactor, in: Proceedings of 2005 International Congress on Advances in Nuclear Power Plants (ICAPP-05), 2005. [17] L. Charpentiera, K. Dawi, M. Balat-Pichelin, E. Beche, F. Audubert, Chemical degradation of SiC/SiC composite for the cladding of gas-cooled fast reactor in case of severe accident scenarios, Corros. Sci. 59 (2012) 127e135. [18] A.R. Raffray, R. Jones, G. Aiello, M. Billone, L. Giancarli, H. Golfier, A. Hasegawa, Y. Katoh, A. Kohyama, S. Nishio, B. Riccardi, M.S. Tillack, Design and material issues for high performance SiCf/SiC-based fusion power cores, Fusion Eng. Des. 55 (1) (2001) 55e95. [19] N.B. Morley, Y. Katoh, S. Malang, B.A. Pint, A.R. Raffray, S. Sharafat,
[20]
[21]
[22]
[23]
[24] [25]
[26]
[27]
[28] [29] [30]
[31] [32] [33] [34]
[35] [36] [37]
S. Smolentsev, G.E. Youngblood, Recent research and development for the dual-coolant blanket concept in the US, Fusion Eng. Des. 83 (7e9) (2008) 920e927. C.P. Deck, G.M. Jacobsen, J. Sheeder, O. Gutierrez, J. Zhang, J. Stone, H.E. Khalifa, C.A. Back, Characterization of SiCeSiC composites for accident tolerant fuel cladding, J. Nucl. Mater. 466 (2015) 667e681. M. Ben-Belgacem, V. Richet, K.A. Terrani, Y. Katoh, L.L. Snead, Thermo-mechanical analysis of LWR SiC/SiC composite cladding, J. Nucl. Mater. 447 (1e3) (2014) 125e142. J.G. Stone, R. Schleicher, C.P. Deck, G.M. Jacobsen, H.E. Khalifa, C.A. Back, Stress analysis and probabilistic assessment of multi-layer SiC-based accident tolerant nuclear fuel cladding, J. Nucl. Mater. 466 (2015) 682e697. Y. Lee, M.S. Kazimi, A structural model for multi-layered ceramic cylinders and its application to silicon carbide cladding of light water reactor fuel, J. Nucl. Mater. 458 (2015) 87e105. L.L. Snead, T. Nozawa, Y. Katoh, T.S. Byun, S. Kondo, D.A. Petti, Handbook of SiC properties for fuel performance modeling, J. Nucl. Mater. 371 (2007) 329e377. R.N. Morris, C.A. Baldwin, R.J. Ellis, J.M. Giaquinto, L.J. Ott, J.L. Peterson, J.E. Schmidlin, 20 GWD SiC Clad Fuel Pin Examination, Oak Ridge National Laboratory, Oak Ridge, TN, 2014. Y. Katoh, T. Nozawa, L.L. Snead, K. Ozawa, H. Tanigawa, Stability of SiC and its composites at high neutron fluence, J. Nucl. Mater. 417 (1e3) (2011) 400e405. A. Campbell, W. Porter, Y. Katoh, L. Snead, Method for analyzing passive silicon carbide thermometry with a continuous dilatometer to determine irradiation temperature, Nucl. Instrum. Methods Phys. Res. B 370 (2016) 49e58. N. Xoubi, R.T. Primm III, Modeling of the High Flux Isotope Reactor Cycle 400, Oak Ridge National Laboratory, Oak Ridge, TN, 2004. High Flux Isotope Reactor Technical Parameters. http://neutrons.ornl.gov/hfir/ parameters. (Accessed 27 July 2016). T. Daly, J. McDuffee, Experimental and Computational Study of the Flux Spectrum in Materials Irradiation Facilities of the High Flux Isotope Reactor, PHYSOR 2012-Advances in Reactor Physics Linking Research, Industry and Education, Knoxville, TN, 2012. L.R. Greenwood, R.K. Smither, SPECTER: Neutron Damage Calculations for Materials Irradiations, Argonne National Laboratory, Argonne, IL, 1985. CINDAS, LLC: Global Benchmark for Critically Evaluated Materials Properties Data. http://cindasdata.com. (Accessed 27 July 2016). MatWeb: Material Property Data. http://matweb.com/. (Accessed 27 July 2016). J.L. McDuffee, Heat transfer through small moveable gas gaps in a multi-body system using the ANSYS finite element software, Minneapolis, in: 2013 Proceedings of the ASME Summer Heat Transfer Conference, 2013. C.V. Madhusudana, Thermal Contact Conductance, Springer-Verlag New York, Inc., New York, NY, 1996. K.J. Negus, M.M. Yovanovich, Correlation of the gap conductance Integral for conforming rough surfaces, J. Thermophys. Heat Transf. 2 (1988) 279e281. S.M.S. Wahid, C.V. Madhusudana, Gap conductance in contact heat transfer, Int. J. Heat Mass Transf. 43 (2000) 4483e4487.
Contents lists available at ScienceDirect
Journal of Nuclear Materials journal homepage: www.elsevier.com/locate/jnucmat
Experimental design and analysis for irradiation of SiC/SiC composite tubes under a prototypic high heat flux* Christian M. Petrie a, *, Takaaki Koyanagi a, Joel L. McDuffee a, Christian P. Deck b, Yutai Katoh a, Kurt A. Terrani a a b
Oak Ridge National Laboratory, P.O. Box 2008, Oak Ridge, TN, 37831, USA General Atomics, P.O. Box 85608, San Diego, CA, 92121, USA
a r t i c l e i n f o
a b s t r a c t
Article history: Received 30 August 2016 Received in revised form 11 April 2017 Accepted 20 April 2017 Available online 4 May 2017
The purpose of this work is to design an irradiation vehicle for testing silicon carbide (SiC) fiberreinforced SiC matrix composite cladding materials under conditions representative of a light water reactor in order to validate thermo-mechanical models of stress states in these materials due to irradiation swelling and differential thermal expansion. The design allows for a constant tube outer surface temperature in the range of 300e350 C under a representative high heat flux (~0.66 MW/m2) during one cycle of irradiation in an un-instrumented “rabbit” capsule in the High Flux Isotope Reactor. An engineered aluminum foil was developed to absorb the expansion of the cladding tubes, due to irradiation swelling, without changing the thermal resistance of the gap between the cladding and irradiation capsule. Finite-element analyses of the capsule were performed, and the models used to calculate thermal contact resistance were validated by out-of-pile testing and post-irradiation examination of the foils and passive SiC thermometry. Six irradiated cladding tubes (both monoliths and composites) were irradiated and subsequently disassembled in a hot cell. The calculated temperatures of passive SiC thermometry inside the capsules showed good agreement with temperatures measured post-irradiation, with two calculated temperatures falling within 10 C of experimental measurements. The success of this design could lead to new opportunities for irradiation applications with materials that suffer from irradiation swelling, creep, or other dimensional changes that can affect the specimen temperature during irradiation. Published by Elsevier B.V.
Keywords: Silicon Carbide Cladding Irradiation Design Composite
1. Introduction Silicon carbide (SiC) is an attractive material for use in nuclear applications due to its radiation tolerance and its strength and chemical inertness at high temperatures [1e4]. In addition to the inherent advantages of SiC, continuous SiC fiber-reinforced, SiC-
* This manuscript has been authored by UT-Battelle, LLC under Contract No. DEAC05-00OR22725 with the U.S. Department of Energy. The United States Government retains and the publisher, by accepting the article for publication, acknowledges that the United States Government retains a non-exclusive, paid-up, irrevocable, worldwide license to publish or reproduce the published form of this manuscript, or allow others to do so, for United States Government purposes. The Department of Energy will provide public access to these results of federally sponsored research in accordance with the DOE Public Access Plan (http://energy. gov/downloads/doe-public-access-plan). * Corresponding author. E-mail address: [email protected] (C.M. Petrie).
http://dx.doi.org/10.1016/j.jnucmat.2017.04.058 0022-3115/Published by Elsevier B.V.
matrix composites (SiC/SiC composites) offer increased damage tolerance, improved mechanical properties, and enhanced reliability compared to monolithic SiC [5e7]. These properties of SiC/ SiC composites make them intriguing candidates for use as structural materials for nuclear applications. One of the specific applications for using SiC/SiC composite materials is their utilization as accident tolerant fuel cladding and/or structural materials in light water reactors (LWRs) [8e13]. Other applications include components in advanced high-temperature gas-cooled reactors (HTGRs) [14], fluoride-salt-cooled high-temperature reactors (FHRs) [15], gas-cooled fast reactors (GFRs) [16,17], and fusion energy applications [2,18,19]. A number of architectures are being investigated for accidenttolerant fuel cladding applications, including fully composite or multilayered architectures incorporating a SiC/SiC composite material and a monolith [6,20]. An important concern is whether the SiC-based cladding is able to maintain hermeticity (and thus fission
C.M. Petrie et al. / Journal of Nuclear Materials 491 (2017) 94e104
product retention). Essentially, the structure is required to withstand the evolving stress state over the fuel lifetime when exposed to intense neutron irradiation under a prototypic high heat flux (and thus a large temperature gradient) as the rod internal pressure increases due to released fission gas buildup. Detailed discussions of the various governing phenomena and the nature of stress evolution in SiC-based cladding structures have been provided elsewhere [21e23]. Briefly, the high heat flux across the thickness of the cladding in LWR applications (~0.5e1 MW/m2) results in a large temperature gradient in SiC/SiC cladding due to the poor thermal conductivity of this composite that further degrades with radiation damage [6]. For metallic cladding, higher thermal conductivity reduces the radial temperature gradient and the associated stresses due to differential thermal expansion across the thickness. Therefore, the primary driver of stress (in the absence of pellet-cladding contact) is the pressure difference between the inside and outside of the cladding. However, in SiC/SiC cladding, the higher radial temperature gradient results in a much more important stress driver that dominates the strain profile across the cladding thickness: the differential swelling strain across the thickness of the cladding. This is the case because swelling is a strong function of temperature in the LWR cladding temperature regime [24]. A previous irradiation experiment investigating fuel-clad interaction with multilayered SiC cladding materials showed cracking near the ends of the cladding where the heat flux (and thus temperature) was lower [25]. Because swelling of SiC is greater at lower temperatures [26], the cladding ends underwent larger volumetric expansion. It is thought that irradiation swelling caused additional stresses as the cladding expanded into the stainless steel capsule housing. These stresses are clearly not truly representative of typical LWR conditions where a somewhat uniform coolant temperature of >300 C is present on the cladding outer surface. This work describes an improved irradiation capsule design aimed specifically at addressing the performance of SiC/SiC cladding in LWR conditions. The design allows for a SiC cladding tube surface temperature in the range of 300e350 C under a representative high heat flux (~0.6 MW/m2) during one cycle of irradiation in an un-instrumented “rabbit” capsule in the High Flux Isotope Reactor (HFIR). Both SiC/SiC composites and chemical vapor deposited (CVD) SiC monoliths have been irradiated, and additional architectures incorporating various combinations of composite and monolith are being considered. To simplify the design and postirradiation examination, the fuel is replaced with a surrogate molybdenum “heater” cylinder that is designed to generate enough heat from gamma heating in the HFIR to achieve the desired heat flux. Design calculations show that the tube surface temperature remains constant to within 12 C despite volumetric irradiation swelling of the tube by as much as 3%. The results of these experiments and future irradiation experiments utilizing the design described in this work will provide experimental validation of fuel performance codes and ultimately improve the fundamental understanding of evolving stress states in SiC/SiC composite materials due to differential thermal expansion and irradiation swelling. Also, specimens irradiated in this configuration are essential for understanding the possibility of microcrack formation in these materials as a result of stress and whether this microcracking will result in release of gaseous fission products. 2. Materials The assembled capsules were designed for insertion into the flux trap of the HFIR (see Section 3.1.1) for irradiation. Neutron and gamma radiation in the flux trap of the HFIR, roughly two orders of magnitude higher than a typical LWR core, cause heating of the
95
capsule materials. Thermal design calculations are performed to estimate capsule component temperatures during irradiation. Temperatures for this capsule are controlled by the geometry, fill gas, and surface roughness of the contacting components. The overall design and assembly of the rabbit capsules is summarized in Fig. 1. Rabbits refer to small sealed aluminum capsules that serve as the irradiation containment vehicle. Fig. 2 shows a picture of the cross-section of an assembled capsule. Molybdenum heater rods (i.e. dense gamma-absorbing cylinders) are inserted inside the cladding tubes. Titanium centering thimbles keep both the cladding and the heaters centered within the capsule housing. A tight-fitting aluminum sleeve is wrapped around the cladding, and an embossed aluminum foil fills the space between the sleeve and the capsule housing. The embossed aluminum foil is the key to this design as it can absorb the irradiation swelling of the cladding without causing significant stress to the cladding. These components make up one sub-assembly. Fig. 1 shows four sub-assemblies: two sub-assemblies that include 16 mm length cladding tubes, one sub-assembly that includes a 12 mm length cladding tube, and one sub-assembly that includes a 4 mm length cladding tube. Alternatively, as was done for the irradiated capsules described in this work, the capsule could be loaded with three sub-assembliesdeach with one 16 mm length cladding tube. A multiple subassembly design allows for multiple cladding tube specimens to be irradiated at once. Grafoil insulating discs (brown components in Fig. 1) are positioned at the top and bottom of the capsule interior to reduce axial heat losses. Table 1 summarizes the capsule components, materials, and nominal dimensions that are relevant to the thermal design. The SiC/SiC tubes were fabricated using a chemical vapor infiltration process at General Atomics. Hi-Nicalon Type S SiC fiber was used for the reinforcement. Two composite tubes with IDs GA-TGI-1 and GA-TGI-4 were investigated. Four monolithic high-purity CVD SiC (Dow Chemical Co.) tubes machined from a plate (IDs CVD-T1, CVDT2, CVD-T6, and CVD-T7) were also investigated. All specimens (both monolith and composite) had inner and outer diameter concentricity within 0.02 mm. Surface roughness was measured to be 0.3 mm root mean square (RMS) for all monolith specimens, 13 mm RMS for GA-TGI-1, and 1.7 mm RMS for GA-TGI-4. The thermometry (or passive temperature monitors) listed in Table 1 are rectangular pieces of CVD SiC positioned inside the molybdenum heaters between the two centering thimbles of a sub-assembly. The thermometry can be examined post-irradiation as a form of passive temperature monitoring to determine the temperature during irradiation [27]. 3. Experimental 3.1. Experimental facilities 3.1.1. HFIR irradiation facility The HFIR is a beryllium-reflected, pressurized, light-watercooled and moderated flux-trap-type reactor located at the Oak Ridge National Laboratory (ORNL). The core consists of aluminumclad involute-fuel plates, which currently utilize highly enriched 235 U fuel at a power level of 85 MW [28]. A typical cycle is 24e25 days. The reactor core, illustrated in Fig. 3, consists of two concentric annular regions, each approximately 61 cm in height. The flux trap is ~12.7 cm in diameter, and the outer fueled region is ~43.5 cm in diameter. The fuel region is surrounded by a beryllium annular reflector, which is in turn backed up by a water reflector of effectively infinite thickness. The reactor core assembly is contained inside a pressure vessel, which is located in a pool of water. The rabbits described in this work are placed in the flux trap of the HFIR in an outer ring target rod rabbit holder (TRRH) position
96
C.M. Petrie et al. / Journal of Nuclear Materials 491 (2017) 94e104
Fig. 1. Exploded computer-aided design view of the SiC cladding tube rabbit assembly.
Fig. 2. Cross-section of an assembled rabbit capsule. For scale, the SiC cladding tube outer diameter is 8.50 mm.
closest to the inner fuel elements. In a TRRH position, up to seven rabbits can be stacked in a target. These rabbits were placed at the reactor midplane (position TRRH-4). The values of the fast (neutron energy > 0.1 MeV) and total neutron flux at this position are 1.1 1015 n/cm2$s and 4.3 1015 n/cm2$s, respectively [30]. Running the SPECTER-ANL code (Argonne National Laboratory) [31] with an average displacements per atom (dpa) cross section of
260 b gives 2.3 dpa in SiC for a 24 day HFIR cycle. 3.1.2. Thermal contact conductance measurement system In order to validate the thermal design simulations (described later in Sections 4 and 5) used to calculate component temperatures during irradiation, experimental measurements of thermal contact conductance through an embossed foil were performed
C.M. Petrie et al. / Journal of Nuclear Materials 491 (2017) 94e104 Table 1 Capsule parts, materials, and nominal dimensions. Part
Material
Dimension
Nominal Value
Housing
Aluminum 6061
Inner diameter Outer diameter
9.52 mm 10.96 mm
Housing end cap Heater
Aluminum 4047 Molybdenum
Cladding
Silicon carbide
Sleeve
Aluminum 1100
Foil
Aluminum 1100
Inner diameter Outer diameter Inner diameter Outer diameter Inner diameter Outer diameter Thickness Depth Spacing
2.50 7.09 7.12 8.50 8.50 9.00 0.05 0.26 0.83
Centering thimble Insulator discs Thermometry
Ti-6Al4V Grafoil Silicon carbide
mm mm mm mm mm mm mm mm mm
97
first pumping down the vessel to ~930 Pa (7 torr) and then backfilling with either helium or neon fill gas. The rabbit capsule design described in this paper utilizes helium fill gas, but both helium and neon gases were tested in the validation experiments in order to test the contact conductance equations used for the capsule design using two different fill gases. While running chilled water through the cooling block, the heater power was ramped up in steps. By measuring the input power and the temperature difference of the surfaces that contact the foil, the contact conductance through the foil, h (W/m2-K), can be calculated as shown in Eq. (1): 00
h¼
q ; DT
(1)
00
Length Cross-section
12.0 mm 1.0 mm 0.5 mm
inside of a sealed, instrumented vessel. Fig. 4 shows a 3D model of the experiment with labeled components. The experiment is placed inside a sealed vessel with additional insulation above and around the perimeter of the experiment to prevent natural convection loops from forming inside the vessel. A cartridge heater is used to provide the heat input. The power to the heater is controlled using a variac and measured using a watt meter. The heat is conducted from the square protrusion of the heating block, through the embossed foil, to the cooling block, which is cooled with chilled (~5 C) water. Thermocouples are inserted into small holes in both the heating block and the cooling block near the surfaces that contact the foil. Additional thermocouples are located in the surrounding insulation to give some indication of heat losses via conduction to the vessel walls. The thermocouples located in the insulation generally read less than 30 C throughout the experiments, indicating that heat losses through the vessel walls to the ambient air are likely minor. 3.2. Thermal contact conductance measurement procedure Thermal contact conductance experiments were performed by
where q (W/m2) and DT are the heat flux through the foil and the temperature difference across the foil, respectively. In this case, DT is the temperature difference of the thermocouples located near 00 the two surfaces in contact with the foil, and q is simply the heater power divided by the cross-sectional area (494 mm2) of the square protrusion of the heating block. 4. Calculations for in reactor performance The thermal design calculations described in this paper determine capsule component temperatures resulting from reactor heating. The ultimate heat sink is the reactor primary coolant, which flows through the reactor core and cools the outer surface of the capsule housing. 4.1. Materials and material properties Material properties and thermal boundary conditions for this work are taken from independently reviewed design and analysis calculations that are maintained internally within ORNL. Material property data are primarily obtained from CINDAS [32], MatWeb [33], and various literature sources. In general, material properties recommended by Snead et al. [24] for CVD SiC were used for all SiC tubes included in the simulations. This includes temperature- and dose-dependent irradiation swelling, temperature-dependent
Fig. 3. Cross-section through the HFIR illustrating the primary experimental sites [29].
98
C.M. Petrie et al. / Journal of Nuclear Materials 491 (2017) 94e104
Fig. 4. Computer-aided design model of experiment for validation of foil contact conductance.
coefficients of thermal expansion, and temperature- and swellingdependent values for elastic modulus and thermal conductivity. Although the elastic modulus for the composite tubes will be different from that of the monoliths, this should not affect the thermal analyses. Thermal conductivity of the irradiated composite tubes was assumed to be 3.0 W/m-K based on previous irradiations of SiC/SiC composites [6]. The nominal swelling of the composite tubes was assumed to be 2.2% (0.73% linear swelling). The effects of varying the irradiation swelling on the thermal contact conductance from the cladding tubes to the capsule housing are discussed later in Section 5.1. No anisotropy was considered in these analyses.
4.2. Heat generation and thermal boundary conditions Peak heat generation rates, parameters for determining axial profiles for the heat generation rates, and convection parameters are summarized in Table 2. The fill gas for this design is helium. The local heat generation rate is estimated with the following profile:
2 z ; qðmaterial; zÞ ¼ qpeak ðmaterialÞ,exp
(2)
s
Where
Table 2 Thermal boundary conditions. Parameter
Value
Convective heat transfer coefficient Bulk coolant temperature Peak heat generation rate for Al-6061 Peak heat generation rate for Ti-6Al4V Peak heat generation rate for SiC Peak heat generation rate for molybdenum Peak heat generation rate for grafoil Correlating parameter (s)
47.1 kW m2 K1 52 C 32.5 W/g 35.2 W/g 32.9 W/g 43.3 W/g 33.7 W/g 30.07 cm
q ¼ local heat generation rate as a function of material and axial location, qpeak ¼ heat generation rate at the HFIR core midplane as a function of material, z ¼ axial location in HFIR, where the core midplane is at zero, and s ¼ correlating parameter.
4.3. Finite-element model Thermal calculations of capsule component temperatures were performed using the ANSYS (ANSYS, Inc.) finite-element software code package. Additional user-written macros were implemented within ANSYS to calculate thermal conductance between contacting parts. These macros account for solid spot conductance, gas gap conductance, and thermal jump distances. Details regarding the solution methods can be found in a previous publication [34]. Thermal contact conductance is extremely important in this particular capsule design, as the high heat flux results in large jumps in temperature at each contact interface. The contact conductance depends on, among other parameters, the thermal conductivity of the fill gas and the surface roughness values for the parts that are in direct contact (i.e., the cladding tube, the sleeve, the foil, and the capsule housing). See Ref. [34] and the references therein (including [35e37]), for details regarding the thermal contact conductance models. The effects of contact pressure on the solid spot conductance and the gas gap conductance are neglected in this analysis as the contact pressures due to clad swelling and foil compression are likely negligible because the foil is very thin (0.05 mm), is very ductile since it is made of a fully annealed 1100 aluminum alloy, and is expected to experience significant irradiation creep. The effects of changing the effective contact area for heat transfer between the foil and the contacting components as the cladding swells and the foil crushes were explicitly modeled. It is later shown in section 5.1 that swelling of the cladding and crushing of the foil reduced the cladding outer surface temperature
C.M. Petrie et al. / Journal of Nuclear Materials 491 (2017) 94e104
by 12 C or less. In order to model the swelling of the cladding and the compression of the embossed aluminum foil, a 2D transverse slice is taken through one sub-assembly inside the rabbit. The 2D model also assumes 10 azimuthal symmetry. A full 3D modeling of foil compression was not attempted due to the small mesh size required to simulate foil compression as the cladding swells. The 2D model is used to determine the effective thermal contact conductance between the cladding and the capsule housing. Then, a 3D model is generated with the foil and sleeve suppressed. The gas gap between the clad and the housing is artificially modified until the contact conductance through this effective gas gap is equal to the calculated contact conductance in the 2D model (see Section 5.1) that includes the sleeve and the compressed aluminum foil. This design is heavily dependent on contact resistance at the clad/sleeve interface, the sleeve/foil interface, and the foil/housing interface. This is why the contact conductance through the embossed foil is validated experimentally. The contact conductance depends significantly on the surface roughness of the contacting surfaces. The surface roughness values of the inner surface of a standard rabbit housing and of an aluminum foil were measured to be 0.26 and 0.65 mm RMS, respectively, using a profilometer. The standard rolls of aluminum foil that are used to fabricate the sleeve have a surface roughness similar to the housing inner surface and the foil (i.e., ~0.3e0.7 mm RMS). The surface roughness of the cladding tube specimens can vary from as low as ~0.3 mm RMS for the CVD monolith specimens to as high as ~13 mm RMS for some composite specimens, as mentioned in Section 2. The surface roughness of the aluminum sleeve was varied in such a way that the effective contact conductance is approximately constant despite variations in clad surface roughness, thus maintaining the desired clad outer surface temperature. All tube specimens are paired with an 8.4 mm RMS sleeve except for composite specimen GA-TGI-1 (13 mm RMS roughness), which uses a 0.3 mm RMS sleeve.
99
The results obtained from the 2D simulations are used to calculate the effective contact conductance from the clad, through the sleeve, through the foil, and to the housing after clad swelling. This is calculated by dividing the heat flux at the outer surface of the clad (0.66 MW/m2) by the change in temperature from the clad outer surface (339 C) to the housing inner surface (68 C). With this method, the contact conductance is calculated to be 2.42 kW/ m2-K. This contact conductance is used in the 3D finite-element simulations (results shown in Section 5.3), which consider the effects of axial heat loss and spatially dependent heat generation. Note that because of the variations in the surface roughness of the cladding specimens and the sleeves, there is some variation in the effective contact conductance from the cladding to the housing. The value of 2.42 kW/m2-K was calculated for the GA-TGI-4 composite tube with an 8.4 mm RMS sleeve. Similar simulations, performed with the GA-TGI-1 composite specimen and with the monolith specimens (and their corresponding sleeves) resulted in contact conductance values of 2.32 and 2.49 kW/m2-K, respectively. To determine the sensitivity of the design temperatures to clad swelling, the linear clad swelling is varied from 0 to 1.05% (from 0 to 3.15% volumetric swelling) for the case with composite tube GATGI-4 and an 8.4 mm RMS sleeve. The 3.15% maximum volumetric swelling that is assumed is the maximum swelling of CVD SiC in the point-defect swelling saturation regime (150e1000 C) [24]. The cladding tube specimens tested in this work were later observed (see Section 6.1) to have linear swelling of up to 0.73% after one 25 day cycle of irradiation in the HFIR. The area-averaged outer surface temperature of the cladding is plotted vs. linear clad swelling in Fig. 6. If the linear swelling was to vary between 0 and 1.05%, the cladding outer surface temperature would be constant to within 12 C. In this scenario, the embossed aluminum foil design is effectively utilized to render the magnitude of the cladding swelling inconsequential to its temperature distribution. 5.2. Experimental validation of thermal contact conductance models
5. Results 5.1. 2D simulation results The 2D simulations are performed to determine the effective contact conductance from the cladding tubes to the capsule housing. Fig. 5 shows temperature contour plots obtained from 2D simulations with composite tube GA-TGI-4 (1.7 mm RMS and an 8.4 mm RMS sleeve). A linear clad swelling of 0.73% (~2.2% volumetric swelling) was assumed for the results shown in Fig. 5. Table 3 summarizes the temperatures of the various components. The heat flux at the outer surface of the cladding was calculated to be 0.655 MW/m2 in the 2D simulations.
Sections 3.1.2 and 3.2 describe the experimental setup and
Table 3 Part temperatures obtained from 2D simulations of composite tube GATGI-4 with 0.73% linear clad swelling. Part
Temperature ( C) Average (MineMax)
Centering thimble Heater Cladding Sleeve Foil Housing
608 (608e609) 601 (596e604) 415 (339e496) 304 (303e305) 135 (131e138) 66 (65e68)
Fig. 5. Contour temperature plot ( C) obtained from 2D simulations of composite tube GA-TGI-4 with 0.73% linear clad swelling.
100
C.M. Petrie et al. / Journal of Nuclear Materials 491 (2017) 94e104
Fig. 6. Composite clad tube GA-TGI-4 outer surface temperature vs. clad swelling determined from 2D simulations.
procedure for the out-of-pile experiments that were performed to validate the thermal contact conductance models used in the thermal design of the rabbit capsule. Figs. 7 and 8 show temperatures near the surface of the heating block (TC1) and the cooling block (TC2), respectively, as well as the heater power as a function of time during experiments with helium (Fig. 7) and neon (Fig. 8) fill gases. It is clear that the temperatures had essentially reached a steady-state for each of the three steps in heater power with both fill gases. Note that the brief decreases in temperature in Fig. 7 during the first 5e10 min were due to cooling of the system with chilled water prior to powering the heater. Using the steady-state temperatures for each step in power in Figs. 7 and 8, a plot of heat flux vs. (TC1 e TC2) or DT can be generated. Performing a linear fit to this data gives contact conductance through the foil for each fill gas. A simple 2D finiteelement model of the experimental rig was constructed to enable a direct comparison of the thermal contact conductance models with the experimental results. Note that the same contact conductance models are used in the experimental validation model and in the rabbit capsule design model. The surface roughness of the heating and cooling blocks was measured to be ~0.95 mm RMS, while the foil roughness was measured to be ~0.63 mm RMS. This information was input in the contact conductance equations in the finite element models. Fig. 9 shows heat flux vs. temperature difference (between the heating and cooling blocks), as determined from the validation
Fig. 7. Temperatures near the surface of the heating block (TC1) and the cooling block (TC2), as well as the heater power, as a function of time during the validation experiments with helium fill gas.
Fig. 8. Temperatures near the surface of the heating block (TC1) and the cooling block (TC2), as well as the heater power, as a function of time during the validation experiments with neon fill gas.
experiments and the finite element models using both helium and neon fill gases. The models show very good agreement with the experimental results. Using error propagation of uncertainties in the thermocouple measurements, power measurements, and dimensional tolerances, uncertainties in the measured values for contact conductance were determined to be ~1e3%. Note that the contact conductance in the validation experiment (~3.9 kW m2 K1 for helium fill gas) differs significantly compared to the contact conductance through the foil in the rabbit irradiation experiment (~2.4 kW m2 K1). This discrepancy is due to differences in surface roughness of the contacting components and the fact that the conductance for the rabbit irradiation capsule includes an extra contact resistance from the cladding/sleeve interface.
5.3. 3D simulation results With the contact conductance models validated via experimental measurements, 3D thermal models of the rabbit capsule were used to determine temperature contours for all of the capsule components. As mentioned in Section 5.1, the 3D models do not include the embossed foil or the sleeve. Instead, an effective gas gap
Fig. 9. Heat flux vs. temperature difference (between the heating and cooling block) data determined from the model and from validation experiments with helium and neon fill gases. Slopes indicate contact conductance through the foil.
C.M. Petrie et al. / Journal of Nuclear Materials 491 (2017) 94e104
between the cladding and the housing was varied manually until the contact conductance from the cladding to the housing was equal to the value determined by the 2D simulations; these 2D simulations do include the foil and sleeve and also account for clad swelling. The effective helium gas gap required to give a contact conductance of ~2.4 kW/m2-K is 83 mm. Fig. 10 shows temperature contour plots obtained from 3D models of the rabbit capsule with composite tube GA-TGI-4 and an 8.4 mm RMS roughness sleeve. The temperature at the outer surface of the tube specimen ranges from 296 to 322 C, with an average temperature of 317 C. Note that the region with relatively low temperature at the top of the tube specimen is due to the fact that the molybdenum heater was designed to be slightly shorter than the tube specimen so that it would not expand more than the tube specimen, which would risk pushing up on the centering thimbles. Therefore, the top of the tube specimen experiences a much lower heat flux. The temperature monitor inside the molybdenum heater has an average temperature of 580 C. This temperature can be compared to the results of post-irradiation dilatometry analyses that determine the actual temperature of the temperature monitor during irradiation. Similar 3D simulations were performed for composite tube GA-TGI-1 and for the CVD monolith tubes. Table 4 summarizes temperature distributions in the tube specimens and temperature monitors for all cases. All simulations resulted in an average clad surface temperature between 310 and 325 C. Temperature gradients through the cladding tubes varied significantly due to differences in the thermal conductivity of the irradiated tubes, which affected the temperatures of the thermometry placed inside the capsule sub-assemblies. The calculated thermometry temperature for the various cladding tubes is discussed in Section 6.2. 6. Discussion The experimental design described in this paper offers a new approach to controlling temperature in uninstrumented rabbit irradiation capsules. The typical approach is to rely on a small (on the order of tens to hundreds of microns, depending on the design) insulating gas gap that requires tight machining tolerances and is sensitive to things such as concentricity misalignment and irradiation swelling. The approach described in this paper relies on surface roughness and the contact area of the embossed foil. It is less sensitive to concentricity misalignment due to the self-
101
centering nature of the foil, and it does not require extremely tight machining tolerances on the outer diameter of the cladding because the foil is designed to crush when the cladding is inserted. This approach is particularly attractive for pressurized creep experiments or other applications where swelling, creep, or other phenomena can affect the dimensional stability of capsule components that comprise an insulating gas gap. 6.1. Post-irradiation examination Initial information regarding the performance of the irradiation capsules utilizing the embossed foil design has been acquired, and a more detailed post-irradiation examination of the tubes will be discussed in a future paper. Of the six specimens that were loaded into the two irradiated capsules (four CVD monoliths and two SiC/ SiC composites), five specimens survived the irradiation and postirradiation cutting operations. The authors believe that the single broken specimen (CVD-T6) was damaged during the postirradiation milling operations required to retrieve the specimens from the capsule housing. Specimen CVD-T2 broke later during post-irradiation handling; it may or may not have also been damaged during post-irradiation disassembly. Fig. 11 shows a picture of one of the rabbit housing cut open by milling, a post-irradiation picture of a monolith tube specimen, and a post-irradiation picture of a composite tube specimen. The periodic lines that appear on the inner surface of the housing indicate that the raised surfaces of the foil remained in contact with the housing as intended. The tube specimens show no signs of surface cracking or other degradation on the outer surface. The shiny sections on the outside of the tubes indicate locations where the aluminum sleeve stuck to the tubes, perhaps during the milling operation to split the capsule housing. Note that specimen GA-TGI1 generally had a very rough surface but it also had a section that was much smoother. These surface variations may have had an effect on the temperatures, which will be discussed in Section 6.2. Dimensional measurements of the cladding tubes were taken before and after irradiation to determine the total diametrical and axial swelling. Table 5 summarizes the irradiation swelling of the tube specimens. Note that specimens CVD-T2 and CVD-T6 were broken, and therefore the post-irradiation length could not be measured. Swelling of the length and outer diameter are mostly consistent, and all of the CVD tubes generally show similar linear swelling (~0.4e0.5%). Composite tube GA-TGI-4 shows similar
Fig. 10. Temperature ( C) contours showing composite clad tube GA-TGI-4 and heater sub-assembly (left), clad only (center), and clad outer surface (right) obtained from 3D simulations.
102
C.M. Petrie et al. / Journal of Nuclear Materials 491 (2017) 94e104 Table 4 Calculated temperatures for each tube specimen and the passive temperature monitors located inside the specimen sub-assemblies. Temperature ( C) Average (MineMax)
Specimen ID
CVD-T1 CVD-T2 CVD-T6 CVD-T7 GA-TGI-1 GA-TGI-4
Specimen
Specimen Outer Surface
Temperature Monitor
334 347 347 334 394 385
310 323 323 310 325 317
470 489 489 470 587 580
(298e367) (308e379) (308e379) (298e367) (304e479) (296e471)
(298e315) (308e325) (308e325) (298e315) (305e329) (297e321)
(469e471) (487e490) (487e490) (469e471) (583e589) (575e581)
Fig. 11. Picture of rabbit housing inner surface after hot cell disassembly (left) and post-irradiation pictures of a monolith tube (center) and composite tube (right).
swelling to the CVD specimens, whereas composite tube GA-TGI-1 is significantly higher (~0.7%). A KEYENCE VHX-1000 optical microscope was used to photograph the embossed foils and determine the profile of the foils before and after irradiation. The disassembly process damaged the foils in certain locations, so information could only be gathered in locations where the tools used in the hot cells did not damage the foil. 3D images were constructed by combining the single images with a different working distance. The step of the working distance was 3 and 10 mm for unirradiated and irradiated specimens, respectively. 2D profiles were obtained by taking an appropriate slice through the 3D image. Fig. 12 shows the 2D profile of a flat embossed foil pre-irradiation and a curved foil post-irradiation. The depth of the foil is approximately 238 ± 4 mm pre-irradiation, and 207 ± 17 mm post-irradiation. The difference in these values (~30 mm) represents the amount that the foil was compressed. This number corresponds well with the observed swelling of the tube specimens (18e30 mm).
6.2. Comparison of predicted vs. measured capsule temperatures Passive SiC temperature monitors were placed inside the molybdenum heaters to give a comparison of the actual irradiation temperature vs. the temperatures predicted from the design calculations. By analyzing the instantaneous coefficient of thermal expansion (CTE) of the temperature monitor during postirradiation isochronal heating and cooling, the median, minimum, and maximum temperatures during irradiation can be determined [27]. Fig. 13 shows the temperatures predicted from the passive thermometers that were placed inside each of the six irradiated tube sub-assemblies compared with the temperatures predicted from the thermal calculations. The experimental values shown are the median temperatures determined from analysis of the postirradiation isochronal annealing data. In general, the predicted temperatures show good agreement with the experimental values. The experimentally determined temperatures inside the four CVD
Table 5 Measured tube specimen irradiation swelling, accurate to within ±0.06%. Material type
Specimen Linear swelling of ID length
Linear swelling of outer diameter
Unirradiated length (mm)
Irradiated length (mm)
Unirradiated outer diameter (mm)
Irradiated outer diameter (mm)
CVD tube CVD tube CVD tube CVD tube SiC/SiC tube SiC/SiC tube
CVD-T1 CVD-T2 CVD-T6 CVD-T7 GA-TGI-1
0.39% N/A N/A 0.51% 0.73%
0.47% 0.55% 0.41% 0.48% 0.70%
16.03 16.03 16.03 16.03 15.95
16.093 N/A N/A 16.112 16.067
8.52 8.52 8.52 8.52 8.57
8.560 8.567 8.555 8.561 8.630
GA-TGI-4 0.46%
0.52%
16.02
16.094
8.54
8.585
C.M. Petrie et al. / Journal of Nuclear Materials 491 (2017) 94e104
103
7. Conclusions
Fig. 12. Measured 2D profiles of a flat unirradiated foil and a curved irradiated foil.
Fig. 13. Temperatures of passive thermometry included inside tube specimen subassemblies, determined from post-irradiation dilatometry, compared with temperatures predicted from thermal calculations.
monolith tubes ranged from 472 to 540 C, while the two composite tube temperatures were 502 and 586 C. The calculated temperature (470 C) for CVD-T1 matches the experimental value to within 2 C. The calculated temperatures for the three remaining CVD tubes are lower than the experimental values by as much as 57 C. For the composite tubes, the calculated temperature (580 C) for GA-TGI-4 matches the experimental value to within 6 C; on the other hand, the predicted temperature (587 C) for GA-TGI-1 is 85 C higher than the experimental value. Given the high heat flux in this design, the temperature at the location of the passive thermometry is sensitive to many variables, including the cladding tube thermal conductivity, tube swelling, and thermal contact conductance between the tubes and the capsule housing. With these considerations, the discrepancies (85 C for the composite tubes and up to 57 C for the monolith tubes) between calculation and experiment are not surprising. Thermal conductivity of the irradiated composite tubes is particularly uncertain given that reported values range from 1.2 to 3.9 W/m-K when the irradiation temperature and dose were in the range of 300e500 C and 0.8e71 dpa, respectively [6]. A more likely explanation is that the relatively smooth section of specimen GA-TGI-1 (see Fig. 11) resulted in increased thermal contact conductance in that section of the specimen, which would result in reduced temperatures. Ongoing work for future publication includes experimental investigations of the irradiation temperature within the SiC tubes based on the dimensional recovery behavior during annealing.
This paper describes a vehicle for testing silicon carbide tube materials under conditions representative of a LWR (constant clad surface temperature of 300e350 C under a high heat flux of ~0.6 MW/m2) in order to validate thermo-mechanical models of stress states in these materials due to differential thermal expansion and irradiation swelling. The design can accommodate clad swelling up to ~3% while keeping the clad surface temperature constant to within 12 C throughout the irradiation. An engineered aluminum foil was developed to absorb the expansion of the cladding due to irradiation swelling. This new approach to controlling temperature (as opposed to the traditional approach utilizing a small insulating gas gap) in uninstrumented rabbit irradiation capsules relies on surface roughness and the contact area of the embossed foil. It is less sensitive to concentricity misalignment due to the self-centering nature of the foil, and it does not require tight machining tolerances on the outer diameter of the cladding because the foil is designed to crush when the cladding is inserted. For the application described in this work, the high heat flux requires the tube specimen to have a relatively constant outer surface roughness in order to prevent variations in temperature due to non-uniform heat transfer from regions of varying surface roughness. This approach is particularly attractive for pressurized creep experiments or other applications where swelling, creep, or other phenomena can affect the dimensional stability of capsule components that would otherwise define an insulating gas gap. A finite-element model of the capsule performance was developed using ANSYS with custom libraries used for calculating thermal contact resistance. The thermal models were validated by out-of-pile testing. Initial post-irradiation examination of the cladding tube specimens reveals that five of the six irradiated tubes survived the irradiation and hot cell disassembly. Inspection of the capsule housing's interior surface and the embossed foils suggests that the design performed as expected. The calculated temperatures of passive SiC temperature monitors placed inside the capsules generally showed good agreement with the temperatures measured post-irradiation via dilatometric analysis, with two calculated temperatures agreeing with experimental values to within less than 10 C. Acknowledgements Lance Snead and Koroush Shirvan at MIT, Greg Markham at Ceramic Tubular Products, and Edgar Lara-Curzio at ORNL provided significant input and constructive criticisms on the irradiation vehicle design. Without the persistent support and encouragement of Frank Goldner at the United States Department of Energy (U.S. DOE) to proceed with this complex design, the completion of this irradiation would not have been possible. Padhraic Mulligan (ORNL) and Kevin Field (ORNL) provided useful comments on the manuscript. Doug Stringfield (ORNL) performed most of the capsule assembly. This research was supported by the Advanced Fuels Campaign of the U.S. DOE, Office of Nuclear Energy. Neutron irradiation in the HFIR was made possible by the Office of Basic Energy Sciences, U.S. DOE. References [1] L.L. Snead, T. Nozawa, M. Ferraris, Y. Katoh, R. Shinavski, M. Sawan, Silicon carbide composites as fusion power reactor structural materials, J. Nucl. Mater. 417 (1e3) (2011) 330e339. [2] L. Giancarli, H. Golfier, S. Nishio, R. Raffray, C. Wong, R. Yamada, Progress in blanket designs using SiCf/SiC composites, Fusion Eng. Des. 61e62 (2002) 307e318.
104
C.M. Petrie et al. / Journal of Nuclear Materials 491 (2017) 94e104
[3] Y. Katoh, L.L. Snead, I. Szlufarska, W.J. Weber, Radiation effects in SiC for nuclear structural applications, Curr. Opin. Solid State Mater. Sci. 16 (3) (2012) 143e152. [4] R. Konings, Comprehensive Nuclear Materials, Elsevier Science, 2011. [5] Y. Katoh, L.L. Snead, C.H. Henager Jr., T. Nozawa, T. Hinoki, A. Ivekovi c, S. Novak, S.M. Gonzalez de Vicente, Current status and recent research achievements in SiC/SiC composites, J. Nucl. Mater. 455 (1e3) (2014) 387e397. [6] Y. Katoh, K. Ozawa, C. Shih, T. Nozawa, R.J. Shinavski, A. Hasegawa, L.L. Snead, Continuous SiC fiber, CVI SiC matrix composites for nuclear applications: properties and irradiation effects, J. Nucl. Mater. 448 (1e3) (2014) 448e476. [7] T. Koyanagi, K. Ozawa, T. Hinoki, K. Shimoda, Y. Katoh, Effects of neutron irradiation on mechanical properties of silicon carbide composites fabricated by nano-infiltration and transient eutectic-phase process, J. Nucl. Mater. 448 (1e3) (2014) 478e486. [8] K. Yueh, K.A. Terrani, Silicon carbide composite for light water reactor fuel assembly applications, J. Nucl. Mater. 448 (1e3) (2014) 380e388. [9] D.M. Carpenter, An Assessment of Silicon Carbide as a Cladding Material for Light Water Reactors, Nuclear Science and Engineering, Massachusetts Institute of Technology, Cambridge, MA, 2011. [10] K.A. Terrani, B.A. Pint, C.M. Parish, C.M. Silva, L.L. Snead, Y. Katoh, Silicon carbide oxidation in steam up to 2 MPa, J. Am. Ceram. Soc. 97 (8) (2014) 2331e2352. [11] Y. Lee, T.J. McKrell, C. Yue, M.S. Kazimi, Safety assessment of SiC cladding oxidation under loss-of-coolant accident conditions in light water reactors, Nucl. Technol. 183 (2) (2013) 210e227. [12] Y. Katoh, K.A. Terrani, Systematic Technology Evaluation Program for SiC/SiC Composite-based Accident-tolerant LWR Fuel Cladding and Core Structures: Revision 2015, Oak Ridge National Laboratory, Oak Ridge, TN, 2015. [13] K.E. Barrett, M.P. Teague, I.J. Van Rooyen, S.M. Bragg-Sitton, K.D. Ellis, C.R. Glass, G.A. Roth, K.M. McHugh, J.E. Garnier, G.W. Griffith, Engineering challenges of LWR advanced fuel cladding technology in preparation for inreactor demonstrations, Trans. Am. Nucl. Soc. (2012) 886e889. [14] L.L. Snead, Y. Katoh, W. Windes, K. Smit, Candidate structural materials for incore VHTR application, Trans. Am. Nucl. Soc. (2008) 1019e1020. [15] C. Forsberg, L. Hu, Design goals for a fluoride-salt-cooled high-temperature test reactor (FHTR), Trans. Am. Nucl. Soc. (2013) 599e602. [16] A. Kohyama, T. Hinoki, T. Mizuno, T. Kunugi, M. Sato, Y. Katoh, J.S. Park, R&D of advanced material systems for reactor core component of gas coolded fast reactor, in: Proceedings of 2005 International Congress on Advances in Nuclear Power Plants (ICAPP-05), 2005. [17] L. Charpentiera, K. Dawi, M. Balat-Pichelin, E. Beche, F. Audubert, Chemical degradation of SiC/SiC composite for the cladding of gas-cooled fast reactor in case of severe accident scenarios, Corros. Sci. 59 (2012) 127e135. [18] A.R. Raffray, R. Jones, G. Aiello, M. Billone, L. Giancarli, H. Golfier, A. Hasegawa, Y. Katoh, A. Kohyama, S. Nishio, B. Riccardi, M.S. Tillack, Design and material issues for high performance SiCf/SiC-based fusion power cores, Fusion Eng. Des. 55 (1) (2001) 55e95. [19] N.B. Morley, Y. Katoh, S. Malang, B.A. Pint, A.R. Raffray, S. Sharafat,
[20]
[21]
[22]
[23]
[24] [25]
[26]
[27]
[28] [29] [30]
[31] [32] [33] [34]
[35] [36] [37]
S. Smolentsev, G.E. Youngblood, Recent research and development for the dual-coolant blanket concept in the US, Fusion Eng. Des. 83 (7e9) (2008) 920e927. C.P. Deck, G.M. Jacobsen, J. Sheeder, O. Gutierrez, J. Zhang, J. Stone, H.E. Khalifa, C.A. Back, Characterization of SiCeSiC composites for accident tolerant fuel cladding, J. Nucl. Mater. 466 (2015) 667e681. M. Ben-Belgacem, V. Richet, K.A. Terrani, Y. Katoh, L.L. Snead, Thermo-mechanical analysis of LWR SiC/SiC composite cladding, J. Nucl. Mater. 447 (1e3) (2014) 125e142. J.G. Stone, R. Schleicher, C.P. Deck, G.M. Jacobsen, H.E. Khalifa, C.A. Back, Stress analysis and probabilistic assessment of multi-layer SiC-based accident tolerant nuclear fuel cladding, J. Nucl. Mater. 466 (2015) 682e697. Y. Lee, M.S. Kazimi, A structural model for multi-layered ceramic cylinders and its application to silicon carbide cladding of light water reactor fuel, J. Nucl. Mater. 458 (2015) 87e105. L.L. Snead, T. Nozawa, Y. Katoh, T.S. Byun, S. Kondo, D.A. Petti, Handbook of SiC properties for fuel performance modeling, J. Nucl. Mater. 371 (2007) 329e377. R.N. Morris, C.A. Baldwin, R.J. Ellis, J.M. Giaquinto, L.J. Ott, J.L. Peterson, J.E. Schmidlin, 20 GWD SiC Clad Fuel Pin Examination, Oak Ridge National Laboratory, Oak Ridge, TN, 2014. Y. Katoh, T. Nozawa, L.L. Snead, K. Ozawa, H. Tanigawa, Stability of SiC and its composites at high neutron fluence, J. Nucl. Mater. 417 (1e3) (2011) 400e405. A. Campbell, W. Porter, Y. Katoh, L. Snead, Method for analyzing passive silicon carbide thermometry with a continuous dilatometer to determine irradiation temperature, Nucl. Instrum. Methods Phys. Res. B 370 (2016) 49e58. N. Xoubi, R.T. Primm III, Modeling of the High Flux Isotope Reactor Cycle 400, Oak Ridge National Laboratory, Oak Ridge, TN, 2004. High Flux Isotope Reactor Technical Parameters. http://neutrons.ornl.gov/hfir/ parameters. (Accessed 27 July 2016). T. Daly, J. McDuffee, Experimental and Computational Study of the Flux Spectrum in Materials Irradiation Facilities of the High Flux Isotope Reactor, PHYSOR 2012-Advances in Reactor Physics Linking Research, Industry and Education, Knoxville, TN, 2012. L.R. Greenwood, R.K. Smither, SPECTER: Neutron Damage Calculations for Materials Irradiations, Argonne National Laboratory, Argonne, IL, 1985. CINDAS, LLC: Global Benchmark for Critically Evaluated Materials Properties Data. http://cindasdata.com. (Accessed 27 July 2016). MatWeb: Material Property Data. http://matweb.com/. (Accessed 27 July 2016). J.L. McDuffee, Heat transfer through small moveable gas gaps in a multi-body system using the ANSYS finite element software, Minneapolis, in: 2013 Proceedings of the ASME Summer Heat Transfer Conference, 2013. C.V. Madhusudana, Thermal Contact Conductance, Springer-Verlag New York, Inc., New York, NY, 1996. K.J. Negus, M.M. Yovanovich, Correlation of the gap conductance Integral for conforming rough surfaces, J. Thermophys. Heat Transf. 2 (1988) 279e281. S.M.S. Wahid, C.V. Madhusudana, Gap conductance in contact heat transfer, Int. J. Heat Mass Transf. 43 (2000) 4483e4487.