PCI analysis of Zircaloy coated clad under LWR steady state and reactor startup operations using BISON fuel performance code

Nuclear Engineering and Design 332 (2018) 383–391

Contents lists available at ScienceDirect

Nuclear Engineering and Design journal homepage: www.elsevier.com/locate/nucengdes

PCI analysis of Zircaloy coated clad under LWR steady state and reactor startup operations using BISON fuel performance code

T

⁎

Nathan Capps , Anh Mai, Michael Kennard, Wenfeng Liu Structural Integrity Associates, 5435 Oberlin Dr, San Diego, Ca 92121, United States

A B S T R A C T

Research following the earthquake and subsequent tsunami that devastated the Fukushima Daiichi nuclear power plant, causing severe damage to the reactor cores, has led to the development of accident tolerant fuel designs. Accident tolerant fuels are described as a fuel form which exhibits improved material response, when compared to traditional uranium dioxide fuel clad by a zirconium alloy, during accident (e.g., loss of coolant accident or reactivity insertion accident) while maintaining or exceeding normal reactor operational expectations. One of the primary goals of accident tolerant fuel concepts is to reduce cladding corrosion by either replacing traditional Zircaloy cladding with a new material or by applying a coating to the outer surface of the cladding with enhanced resistance to corrosion at high temperatures. However, while coating the clad may provide an increase in high temperature corrosion resistance, it also changes the mechanical state of the cladding and potentially creates alternate failure mechanisms during reactor operations or accident conditions. This research has primarily focused on the pellet-cladding interactionmechanical response of FeCrAl (Iron-ChromiumAluminum alloy) coated cladding during light water reactor steady state and reactor startup operations. Using the MOOSE-based, finite-element fuel performance code BISON and commercial pressurized water reactor data obtained from Diablo Canyon Unit 2, a preliminary analysis was performed to evaluate the performance of FeCrAl coated cladding under steady state operation and its mechanical stability under pellet-cladding interaction during a reactor startup. This work summarizes BISON 2-D radial-axially symmetric (R-Z) and radialcircumferentially symmetric (R-θ) simulation results by comparing 20, 50, and 100-µm FeCrAl coating thicknesses as well as FeCrAl cladding to traditional Zr-4 cladding under pellet-cladding interaction condition. The results of this study will provide a feasibility analysis for steady state and transient operational response of FeCrAl coated, Zircaloy cladding. Furthermore, this work will provide a fundamental basis for the assessment of the ability of coated cladding, as an ATF candidate, to maintain or exceed current steady state reactor operating expectations.

1. Introduction The dominate fuel design in the current U.S light water reactor (LWR) fleet utilizes uranium dioxide (UO2) clad in a zirconium alloy. Years of research accompanied by 60 years of reactor operational experience has steadily generated technological advancements as well as an extensive experience base of both material response and commercial and test reactor data, under steady state and transient conditions. This has lead current fuel technologies to reach the limit of their performance capabilities. The U.S. Department of Energy (DOE) along with the commercial nuclear energy industry has placed an emphasis on developing fuels designed to reach higher burnup specifically for the minimization of nuclear waste, increasing fuel power density, and improving fuel reliability (Carmack et al., 2013; Kiran Kumar et al., 2016;

⁎

Stachowski et al., 2016). The events at the Fukushima Daiichi nuclear power generating station coupled with recent material advancements have provided a strong motivation for the industry and regulatory agencies to improve nuclear fuel performance, safety, and reliability by developing new fuel designs to address the inherent vulnerabilities of current fuel form during beyond design basis accidents, while maintaining current operational standards. To address these vulnerabilities, the U.S. DOE has invested in developing the next generation of nuclear fuel with the specific goal of improving fuel rod response while increasing its ability to tolerate a severe accident for a considerably longer period of time, as compared to conventional UO2-Zircaloy fuel forms (Carmack et al., 2013; Kiran Kumar et al., 2016; Stachowski et al., 2016). Accident tolerant fuels (ATFs) are designed, in essence, to increase

Corresponding author. E-mail address: [email protected] (N. Capps).

https://doi.org/10.1016/j.nucengdes.2018.03.045 Received 15 November 2017; Received in revised form 29 March 2018; Accepted 31 March 2018 0029-5493/ © 2018 Published by Elsevier B.V.

Nuclear Engineering and Design 332 (2018) 383–391

N. Capps et al.

2016). During a local power increase from a condition of reduced or closed pellet-cladding gap, pellet expansion produces high contact forces between the fuel pellet and the cladding, and during rapid thermal expansion of the pellet, the fuel cracks open further, inducing tangential shear forces on the cladding inner surface. These tangential shear forces are a function of the pellet-cladding gap residual contact pressure prior to a power maneuver, power level at gap closure, interfacial friction, and the maximum local power. The addition of a coating changes the mechanics governing PCI, first, by adding an exterior coating foreign to the parent material, and secondly, by reducing the Zircaloy thickness in order to maintain consistent fuel rod geometry characteristics. An outer surface FeCrAl coating has vastly different material response than Zircaloy. Under PCI conditions, the coated cladding will be exposed to temperature changes and mechanical load induced by the fuel pellet expansion. Unlike a uniform single cladding material, the coated cladding will need to be treated as a composite, accounting for each constituent’s thermal expansion, elastic moduli, irradiation and thermal creep, irradiation hardening, irradiation growth and/or swelling, and plastic deformation models. In order to accomplish this complex analysis, a fuel performance code will need to be geometrically flexible as well as be able to handle complex material models, the coupling between each material model, and interaction between the two constituents in the system. Fuel performance codes capable of modeling these failure mechanisms are multi-dimensional finite-element codes. The BISON fuel performance code provides unique material and physical behavior modeling capabilities as well as its finite-element versatility of spatial geometric representations, and for these reasons, BISON will be used to better understand these coated cladding concepts under PCI conditions. BISON is built upon the Multi-Physics Object-Oriented Simulation Environment (MOOSE) developed at Idaho National Laboratory (INL) (Capps et al., 2016). MOOSE is a parallel finite element computational system that uses a Jacobian-free, Newton-Krylov (JFNK) method to solve coupled systems of non-linear partial differential equations (Capps et al., 2016). Using advanced fuel rod modeling capabilities in BISON, this study will provide a foundational analysis to understand the underlying changes in mechanical stability and provide fuel designers and engineers with a fundamental understanding of the performance of coated clad, as an ATF candidate.

coping time following a beyond design basis accident scenario and should also meet requirements for design basis accident conditions, such as a loss of coolant accident (LOCA) or reactivity insertion accident (RIA), while preserving or improving current steady state reactor operational performance (Carmack et al., 2013; Kiran Kumar et al., 2016; Stachowski et al., 2016). In order to demonstrate the improved fuel rod response, the advanced fuel form must reduce cladding oxidation reaction kinetics, minimize hydrogen generation rate, improve cladding thermo-mechanical properties, and reduce fission product release (Carmack et al., 2013; Kiran Kumar et al., 2016; Stachowski et al., 2016). While fission gas release is primarily controlled by the thermomechanical state of the fuel pellet, hydrogen generation and steam kinetics phenomena are primarily driven by the cladding material exposed to the coolant. When exposed to high temperature steam, the steam interaction with zirconium-based cladding alloys results in an exothermic oxidation reaction. Furthermore, the rapid oxidation of the cladding produces large amounts of hydrogen coupled with the exothermic oxidation reaction, can lead to an explosion such as the one witnessed at the Fukushima Daiichi Unit 3. In addition to the increased reaction kinetics, the oxidation and hydride formations severely deteriorates the clad structural integrity which could lead to a cladding breach and subsequent release of fission products. One of the primary goals for ATF concepts is to minimize high temperature steam-clad reaction kinetics and ensuing, hydrogen generation, by either replacing the Zircaloy cladding with a new material or by applying a coating on the outer surface of the cladding with enhanced corrosion resistance at high temperatures. One proposed concept applies an iron-chromiumaluminum alloy (FeCrAl) coating on the cladding outer surface to improve the high temperature corrosion resistance (Carmack et al., 2013; Kiran Kumar et al., 2016). Under steady state conditions, FeCrAl offers an improved high temperature corrosion response similar to other stainless steels used inside the reactor core. Furthermore, FeCrAl being a stiffer material than traditional cladding types, may assist in mitigating fuel failure that have plagued the industry, such as debris and fretting (Stachowski et al., 2016; Rebak et al. 2016; Cox, 1990). Under accident scenarios, FeCrAl has the ability to form both chromium oxide and aluminum oxide, depending on the temperature regime, dramatically improving the materials resistance to superheated steam, in turn reducing the hydrogen generation rate and maintaining a coolable geometry for longer periods of time (Stachowski et al., 2016; Rebak et al. 2016; Cox, 1990). However, the base material of FeCrAl being iron is less transparent to neutrons than zirconium alloys, having an impact on fuel cycle and electricity generation cost. Furthermore, the release of tritium into the coolant is of increased concern for iron-based alloys. However, the addition of a coating to the cladding outer surface will help alleviate the neutron penalty associated with FeCrAl alloys while improving the cladding mechanical response under normal operating conditions, specifically pellet-clad interaction (PCI). PCI in LWR fuel is a coupled thermal-chemical-mechanical process that can lead to a breach in the cladding and subsequent release of radioactive fission products into the coolant under certain conditions of operating history, power changes, and fuel rod design characteristics (Lyons et al., 1963; Roberts and Gelhaus, 1979; Garzarolli et al., 1979; Gaston et al., 2009). Reactor operating restrictions, which limit power maneuvering, have been utilized to mitigate PCI, however, they restrain operational flexibility, leading to a loss of power generation. PCI failures generally occur in previously irradiated fuel subsequent to a rapid increase in local power after a period of time at a lower or reduced power level. Pellet Cladding Interaction Stress Corrosion Cracking (PCISCC) failure is driven by localized stress in the vicinity of a pellet crack, combined with the presence of aggressive chemical species, such as iodine and cesium, which induces stress-corrosion-cracking (SCC) in the cladding (Montgomery et al., 2013; Williamson et al., 2016). Radial fuel pellet cracks form in brittle ceramic pellets as a result of large radial temperature gradients and are believed to be an important factor in the PCI failure mechanism (Montgomery et al., 2013; Williamson et al.,

2. PCI modeling approach The PCI modeling approach used in this article has been described in detail in Ref (Montgomery et al., 2013; Williamson et al., 2016; Rashid et al., 1988) and closely follows the methodology developed in Reference (Groeschel et al., 2002; Sunderland et al., 1999; Nesbit et al., 2009; Capps et al., 2016). The fuel rod analysis effort consists of two main steps, which together are used in order to determine the fuel rod characteristics following a base irradiation as well as the fuel rods mechanical response under PCI conditions. Step 1 consists of a steady state R-Z depletion calculation for the first cycle of operation, which establishes the fuel rod steady state conditions: pellet-cladding gap, plenum pressure, and released fission gas, following the first cycle of operation. The results of this analysis provided input conditions for the local-effects analysis step. Additionally, this assessment consists of a full-length R-Z analysis of the second cycle startup immediately following the first cycle of operations. The purpose of this analysis is to identify the axial location in the fuel rod where the maximum cladding hoop stress occurs. The second step is the initial local effects simulations which uses output parameters from the R-Z simulations in Step 1 defined by the location in Step 2. The data is utilized as input for the 2-D radial-circumferential (R-θ) or full 3-D geometric model corresponding to the axial location identified from the previous step. A pictorial description of this process can be seen in Fig. 1 384

Nuclear Engineering and Design 332 (2018) 383–391

N. Capps et al.

Fig. 1. PCI modeling methodology used to assess the fuel rod state following an initial cycle of irradiation and to understand the localized effects on cladding performance due to pellet cracking (EPRI Report, 1999).

3. Diablo Canyon Unit 2 Cycle 8 and Cycle 9 startup

Table 2 Description of the reactor coolant conditions during both cycles of irradiation (EPRI Report, 1992).

Diablo Canyon Unit 2 is a 4-loop Westinghouse designed PWR with an operating core power of 3411 MWth (EPRI Report, 1992). The reactor core consists of 193 Westinghouse manufactured 17 × 17 fuel assemblies containing UO2 fuel clad in Zircaloy-4. Westinghouse has designed specific restart ramp rate restrictions for Diablo Canyon Unit 1 and 2 in order to reduce the potential for PCI induced failure. The imposed ramp rate following a refueling outage consisting of a 30%/h ramp up to a specified threshold of 20% of full reactor power. Once the threshold power is reached, the linear heat rate is further limited to 3% per hour until full power is reached (EPRI Report, 1992). Although this startup recommendation may be out of date, the purpose of this article is to assess the mechanical response of FeCrAl coated Zr-4 under prototypical PWR startup conditions. Therefore, Tables 1 and 2 summarize the pre-irradiation characteristics for Westinghouse manufactured 17 × 17 fuel rods, as well as reactor coolant conditions for Diablo Canyon Unit 2. This information is paramount to the development of realistic PWR fuel rod conditions used as input to BISON.

Value

Units

Fuel Stack Height Cladding Length Cladding Outer Diameter Cladding Inner Diameter Cladding Type Internal Gas Pressure (at STP) Gas Type Fuel Pellet Outer Diameter Fuel Pellet Length Fuel Enrichment Fuel Grain Size Initial Density Heat Deposited in Fuel

144 151.6 0.36 0.315 Zr-4 350 He 0.3088 0.370 3.6 22.0 95.67 0.974

inches inches inches inches

Value

Units

Inlet Temperature Pressure Mass Velocity Subchannel Equivalent Hydraulic Diameter

540.7 2250 2.26 0.510

°F psia Mlbm/ft2/sec in

Both the Cycle 8 core axial and radial power distributions were constructed by Pacific Gas and Electric (PG&E) based on ANC simulations and extracted from Ref. (EPRI Report, 1992). This information is vital for the establishment of the fuel rod conditions during the refueling outage prior to the Cycle 9 startup. The PG&E provided power history was generated as a function of core average burnup, where the axial and assembly peaking factors were extracted from the ANC calculation and applied to the core average power. In order to provide a reasonable fast neutron flux multiplier, data extracted from the North Anna plant was used (Seiler et al., 2011). The fast flux multiplier used for the Diablo Canyon Unit 2 power history calculation is 1.6n/cm2 per kW/ft. Cycle 9 startup utilizes Westinghouse prescribed ramp rate restrictions, restricting the initial ramp rate to 30%/hr to 20% threshold power and 3%/hr above the 20% until full power. Illustrated in Fig. 2a are a variety of Cycle 8 rod average heat rates used to evaluate the coated clad and FeCrAl concepts dependence on burnup. Secondly, Fig. 2b illustrates the local linear heat rate (LHR) for Cycle 9 startup based on Westinghouse ramp rate restrictions as described earlier. Power histories from Fig. 2a will be used to test both fuel and cladding mechanical responses as well as fission gas release as a function of coating thickness and cladding type. However, the original power history for this rod, extracted from Ref (EPRI Report, 1999), corresponded to the highest power rod in the core. Typical PCI-induced failures are known to occur in fuel rods with a cycle of irradiation ranging between 24 and 34 GWd/tU. This article assesses the effects of fuel rods with average burnups in the range of 31–37 GWd/tU. The local power,

Table 1 Pre-irradiated characterization data for Westinghouse manufactured 17 × 17 fuel rods (EPRI Report, 1992). Fuel Rod Parameter

Coolant Channel Parameter

psia inches inches % micron % TD –

385

Nuclear Engineering and Design 332 (2018) 383–391

N. Capps et al. 30

45 40

25

Linear Heat Rate (kW/m)

Linear Heat Rate (kW/m)

35 20

15

10

37.3 GWd/tU 36.1 GWd/tU 35 GWd/tU 33.9 GWd/tU 32.8 GWd/tU 31.8 GWd/tU

5

30 25 20 15 10 5 0

0 0

100

200

300

400

500

600

14760 14770 14780 14790 14800 14810 14820 14830 14840 14850

700

Time (hours)

Time (Days)

(a)

(b)

Fig. 2. Diablo Canyon Unit 2 Cycle 8 rod average and peak linear heat rates for a) steady state power history for once burned fuel and b) LHR for Cycle 9 startup based on Westinghouse ramp rate restrictions (EPRI Report, 1992).

shown in Fig. 2b, is indicative of a localized power increase consistent with a typical aggressive reactor startup.

Table 3 List of test cases used for this analysis along with the corresponding Zr-4 and Coating thicknesses.

4. Results and discussion 4.1. FeCrAl coated cladding The objective of this research is to understand the fuel rod mechanical response due to the presence of an outer surface coating to the cladding, to assess the fuel rod performance under a PCI startup conditions, and to conduct a sensitivity analysis by varying the coating thickness. In order to accomplish this, fuel rod and reactor characterization data were extracted from Reference (EPRI Report, 1992) and used to simulate steady state PWR and reactor startup conditions. Secondly, full length R-Z fuel rods and subsequently local effect R-θ geometric meshes were constructed for a number of coating thicknesses used for this sensitivity analysis. Fig. 3 illustrates a full-length R-Z mesh and the corresponding R-θ mesh for the case of cladding (Zr-4) thickness of 461 µm with a 100 µm FeCrAl coating on the outer surface (the FeCrAl layer is shown in red). Similar models were constructed for each case analyzed. Table 3 lists each case along with the thicknesses of Zr-4 substrate and the FeCrAl coating thickness, used for the analysis. Note

Case Number

Fuel Type

Zr-4 Thickness (micron)

FeCrAl Coating Thickness (micron)

1 2 3 4 5

UO2 UO2 UO2 UO2 UO2

561 541 511 461 0

0 20 50 100 561

that the total cladding thickness in each case remains constant at 561 µm. In order to understand how the coating will respond under startup PCI conditions, it is important to understand the fuel rod history prior to reactor startup. Fig. 4 compares the cladding creep down response for each cased documented in Table 3. First looking at the “no coating” case, e.g. standard Zr-4 fuel rod, the cladding creeps down rather quickly until pellet-cladding contact occurs, at which point the fuel pellet drives the cladding radially outward until the end of the cycle. Looking next at the simulations containing a FeCrAl coating, a distinct trend emerges. As the FeCrAl coating thickness increases, cladding thermal expansion increases and cladding creep down decreases. Having a composite material, the global material response is directly impacted by each material, and as the secondary material (i.e., the FeCrAl coating in this case) thickness increases, the response shifts away from the parent material (i.e., Zr-4), toward the fuel rod response of the FeCrAl cladding. Cladding creep down response is modeled similarly. FeCrAl being a much stiffer material is less susceptible to both thermal and irradiation creep, which can be observed by looking at the FeCrAl cladding cases (Terrani et al. (2016)). With Zr-4 as the parent material, cladding response will be representative of the Zr-4 material response, however, as the FeCrAl coating thickness increases and the corresponding Zr-4 thickness decreases, the cladding creep down is slowed and begins responding in a manner similar to the FeCrAl cladding case. However, there is one issue with the current results. BISON, currently, lacks an irradiation hardening model for FeCrAl. Irradiation hardening would increase the yield stress and stiffness, resulting in changes to the cladding deformation and stress state. Once the pelletcladding gap closes, fuel swelling drives cladding deformation until the end of the cycle, as seen in Fig. 4b. Shifting the focus to Fig. 4b, as the coating thickness increases the

Fig. 3. a) Full length R-Z fuel rod and b) R-θ geometric meshes representations for a 100-µm FeCrAl coating. 386

Nuclear Engineering and Design 332 (2018) 383–391

N. Capps et al. 100

Fuel Radial Displacement (microns)

Clad Radial Displacement (microns)

30

20

10

0

No Coating 20 microns 50 microns 100 microns 561 microns

-10

90

80

70

60 No Coating 20 microns 50 microns 100 microns 561 microns

50

40

-20 0

2000

4000

6000

8000

10000

12000

14000

16000

0

2000

4000

6000

8000

10000

12000

14000

16000

Time (hr)

Time (hours)

(a)

(b)

Fig. 4. Cycle 8 steady state time history comparison of a) cladding radial displacement and b) fuel radial displacement as a function of FeCrAl coating thickness for a final rod average burnup of 37.3 GWd/tU.

difference between the fuel and cladding radial displacements are comparable, resulting in a similar gap size. As the FeCrAl coating thickness increases, cladding creep-down decreases and fuel temperatures begin to increase, as a result of the larger pellet-cladding gap. While under irradiation, both solid and gaseous fission products are being produced, and the gaseous fission products form both intergranular and intragranular bubbles. Under irradiation, as the amount of gaseous fission products increases, both the internal pressure and volume of the bubble may increase, depending on the mechanical state of the fuel. For the FeCrAl cladding case, the pellet-cladding gap remains open for the majority of the simulation with the fuel pellets operating at higher temperatures, leading to the formation of larger fission gas bubbles in the fuel. As the rod transitions from full power to HZP, the fuel temperature will decrease directly affecting the bubbles pressure, while maintaining constant bubble volume. In essence, initial fuel temperatures are higher for both zirconium alloy coated with FeCrAl and FeCrAl cladding resulting in an increase in gaseous swelling. The increase in gaseous swelling increases the fuel pellet radial displacement, and the increase in fuel radial displacement trend holds for all evaluated burnups. The FeCrAl cladding case shows a vastly different pellet-cladding gap and cladding radial displacement trend, as compared to the rest of the cladding types. As the rod average burnup decreases, the pellet-cladding gap increases, while the cladding radial displacement shows a negligible difference. This suggests the pelletcladding gap for a fuel rod cladding in FeCrAl is strongly dependent on the fuel pellet radial expansion and not dependent on cladding creep down. Further evidence is shown by looking at the FeCrAl cladding creep down as a function of burnup. As shown in Fig. 6b, the change in cladding radial displacement from 37.3 to 31.8 GWd/tU is less than 5 µm, while the fuel radial displacement, shown in Fig. 6a, follows a similar trend as the simulations using FeCrAl coating or traditional Zircaloy clad. Lastly, fission gas release is an important concern for a PCI induced by stress corrosion cracking (SCC). Looking at Fig. 7, as both the coating thickness and rod average burnup increase, the amount of fission gas released into the pellet-cladding gap and plenum increases. With an increased fission gas inventory in the pellet-cladding gap, the potential for SCC increases as well. Furthermore, the fission gas release also reduces the pellet-cladding gap conductance thereby increasing temperatures, pellet thermal expansion, and subsequently cladding stresses. By increasing both the fission gas inventory and, potentially, the cladding stress, SCC will inherently be increased. This phenomenon provides a point of concern moving forward with coated claddings. Fig. 8 shows the sensitivity of the cladding inner and outer surface

fuel radial displacements likewise increase. With both the cladding initially expanding more and creeping down less, the pellet-cladding gap is larger. Gap thermal conductance has a strong dependence on the pellet-cladding gap. Larger pellet-cladding gaps result in lower thermal heat transfer and higher fuel temperatures; leading to an increase in the pellet thermal expansion. Once the pellet-clad gap closes the fuel temperatures remain relatively constant, and fuel swelling drives any further pellet displacements. Following the Cycle 8 base irradiation, a residual pellet-cladding gap remains, which primarily depends on differential thermal expansions, cladding creep, and plenum pressure. Fig. 5 compares BISON pellet-cladding gap predictions at hot zero power (HZP) condition as a function of coating thickness and rod average burnup. Following gap closure, pellet swelling drives any further cladding deformation by pushing, radially outwards, on the cladding, moreover, pellet radial displacement has a dramatic impact on the residual pellet-cladding gap. As seen in Fig. 5, the pellet-cladding gap for all five cladding types behave similarly, with the largest differences observed for the fully Zr-4 and FeCrAl (561 µm) cladding simulations. In order to determine the cause of this response, Fig. 6 compares post Cycle 8 irradiation fuel and cladding radial displacement results at HZP. Both Zr-4 and FeCrAl have different fuel and cladding radial displacements, however, the

Residual Pellet-Clad Gap at HZP (microns)

26 37.3 GWd/tU 36.1 GWd/tU 35 GWd/tU 33.9 GWd/tU 32.8 GWd/tU 31.8 GWd/tU

24 22 20 18 16 14 12 10 8 0

100

200

300

400

500

600

FeCrAl Coating Thickness (microns)

Fig. 5. Post Cycle 8 steady state irradiation pellet-cladding gaps as a function of coating thickness at HZP. 387

Nuclear Engineering and Design 332 (2018) 383–391

N. Capps et al. 15

Clad Radial Displacement at HZP (microns)

Fuel Radial Displacement at HZP (microns)

75

70

65

60 37.3 GWd/tU 36.1 GWd/tU 35 GWd/tU 33.9 GWd/tU 32.8 GWd/tU 31.8 GWd/tU

55

50

10

5

0

-5

37.3 GWd/tU 36.1 GWd/tU 35 GWd/tU 33.9 GWd/tU 32.8 GWd/tU 31.8 GWd/tU

-10

-15 0

100

200

300

400

500

600

0

100

FeCrAl Coating Thickness (microns)

200

300

400

500

600

FeCrAl Coating Thickness (microns)

(a)

(b)

Fig. 6. HZP a) fuel and b) cladding radial displacement following Cycle 8 base irradiation for the six rod average burnups. 2.0

18

1.8

16

Pellet-Clad Gap (microns)

Fission Gas Release (%)

1.6 1.4 1.2 1.0 0.8 37.3 GWd/tU 36.1 GWd/tU 35 GWd/tU 33.9 GWd/tU 32.8 GWd/tU 31.8 GWd/tU

0.6 0.4 0.2

20 microns 50 microns 100 microns 0 microns

14 12 10 8 6 4 2

0.0

0 0

100

200

300

400

500

600

0

1

2

3

4

FeCrAl Coating Thickness (microns)

Outer Surface Peak Cladding Hoop Stress (MPa)

Inner Surface Peak Cladding Hoop Stress (MPa)

380

370

360

350

340 FeCrAl Plasticity off FeCrAl Plasticity on 330 100

150

200

250

300

350

400

7

8

9

10

Fig. 9. BISON R-θ Cycle 9 reactor startup time history results for a rod average burnup of 37.3 GWd/tU comparing FeCrAl coating thickness pellet-cladding gap to traditional Zircaloy cladding.

390

50

6

Time (hr)

Fig. 7. BISON fission gas release results as a function of FeCrAl coating thickness and rod average burnup.

0

5

450

500

550

600

430 FeCrAl Plasticity off FeCrAl Plasticity on

420 410 400 390 380 370 360 350 0

50

100

150

200

250

300

350

400

450

500

550

600

FeCrAl Coating Thickness (microns)

FeCrAl Coating Thickness (microns)

(a)

(b)

Fig. 8. R-θ BISON results comparing clad peak hoop stress on a) inner surface and b) outer surface of the clad for a rod average burnup of 37.3 GWd/tU as a function of coating thickness with and without FeCrAl plasticity model. 388

Nuclear Engineering and Design 332 (2018) 383–391

N. Capps et al.

reached, which correlates to reaching full power. As discussed earlier, irradiation hardening of FeCrAl is not currently modeled. This simplification results in higher than expected FeCrAl creep rates and lower than expected peak hoop stress values. Future work should include the development and implementation of a FeCrAl irradiation hardening model to better represent the FeCrAl mechanical response under PCI conditions. Considering the burnup effect on cladding hoop stress for the Zr-4 case, observed in Fig. 11, as burnup decreases, generally, the inner and outer surface hoop stress increase. This can be attributed to the pelletcladding gap thickness. However, the trend for the coated cladding, illustrated in Fig. 11a, shows the hoop stress for both inner and outer clad surfaces increasing as burnup decreases for all cases. The difference seems to be a result of the change in cladding stiffness coupled with the change in pellet-cladding gap. The addition of a FeCrAl coating to the cladding outer surface increases the cladding stiffness resulting in the pellet-cladding gap closing later during the startup, as shown in Fig. 9. Coupling the stiffer composite material with a smaller pelletcladding gap inherently increases contact forces, which in turn increases the imposed hoop stress on the cladding inner surface. However, the trend of hoop stress decreasing as a function of coating thickness remains for all cases, further illustrating the hoop stress dependence on the cladding material properties. As for the FeCrAl clad case and referring back to Fig. 5, as burnup decreases the pellet-cladding gap becomes increasingly larger. Once the pellet-cladding gap is sufficiently large, the pellet-cladding contact will occur late in the reactor startup, and the hoop stresses will be reduced significantly as seen in Fig. 11a. Lastly, results shown in Fig. 11b describes the burnup dependence on the outer surface hoop stress. As for the cases with a FeCrAl coating, the hoop stress approaches the yield stress for unirradiated FeCrAl (∼375–425 MPa) at operating temperature conditions. As the FeCrAl hoop stress begins to exceed the yield stress, the material will begin to yield, producing a plastic strain and plastic strain hardening. FeCrAl plastic strain results as a function of burnup and coating thickness can be seen in Fig. 12. Both Figs. 11 and 12 illustrates that lower burnups produce both higher stresses and strains. As noted earlier, PCI induced failure is most common in fuel rods that have undergone a cycle of irradiation, accumulating a rod average burnup between 24 and 34 GWd/tU. As the burnups assessed here are > 31 GWd/tU, future work should consider a wider range of accumulated rod average burnup. Furthermore, BISON does not have a FeCrAl irradiation hardening model, nor does BISON have the capability to translate plastic strain hardening from the R-Z base irradiation to the local effects

hoop stress to the FeCrAl layer thickness with an elastic material model and a plasticity model. Fig. 8a shows that the uncoated Zr-4 case has a higher inner surface hoop stress than any of the coated cladding cases. As the FeCrAl coating thickness increases, the inner surface hoop stress decreases. Note that FeCrAl irradiation hardening and any plastic strain hardening accumulated during the base irradiation is currently being neglected. However, the trend of the dependency on the coating thickness is expected to be similar with different hardening models. The difference between the hoop stress result for each case is related to small differences in pellet-cladding gap response, shown in Fig. 9. As the thickness of the FeCrAl coating increases, pellet-cladding gap closure occurs at a later time during the ramp. Considering Fig. 9, the 20µm coating case illustrates the cladding creeping down slightly faster than the other cases. This is related to the composite nature of the coated cladding. FeCrAl being a stiffer material than Zr-4, increases the global cladding stiffness, inherently reducing the ability for the cladding to creep down. Therefore, as thickness of the FeCrAl coating increases, the pellet-cladding gap closes later in the ramp and reduces the hoop stress on the cladding inner surface. Returning back to Fig. 8 and comparing the elastic to the elastic-plastic model results, the elasticplastic cases show an increase in the inner surface hoop stress and a decrease in the outer surface hoop stress. As the FeCrAl deforms plastically, it also becomes stiffer due to plastic strain hardening. Increasing the FeCrAl stiffness reduces the cladding mechanical response, and in turn increases the inner surface hoop stress. Note that accumulated plastic deformation during the R-Z base irradiation case was not accounted for in the local effects simulations. Future work should include applying a non-zero plastic strain to the local effects cases. The cladding inner surface hoop stress is expected to be higher as a result of the stiffer coating. Future work would need to include plastic strain hardening from the R-Z base irradiation in the local effects models. However, when making a direct comparison between Zr-4 and FeCrAl cladding materials, BISON predicts higher hoop stress in the FeCrAl cladding, illustrated in Fig. 8a as well as the hoop stress time history during the reactor startup shown in Fig. 10a. Once the pelletcladding contact occurs, FeCrAl being a stiffer material, even without irradiation hardening, the use of a stiffer material, such as FeCrAl, results in higher imposed stresses than for Zr-4 alone. This can also be seen in Fig. 8b, where the hoop stress on the cladding outer surface is higher in the models containing FeCrAl coatings. This trend, however, will only hold while the pellet-cladding gap for both models are similar. Once the FeCrAl gap becomes larger, the inner and outer surface hoop stresses will decrease drastically. Another important observation is related to the rate of stress relaxation once the peak hoop stress has been

400

Outer Surface Clad Hoop Stress (MPa)

Inner Surface Clad Hoop Stress (MPa)

500

400

300

200

100

0

-100

Zr-4 FeCrAl

-200

300

200

100

0

-100 Zr-4 FeCrAl -200

0

10

20

30

40

50

60

70

80

90

100

0

10

20

30

40

50

Time (hr)

Time (hr)

(a)

(b)

60

70

80

90

100

Fig. 10. Time history hoop stress plot of the a) inner and a) outer surface of an R-θ Zr-4 and FeCrAl cladding material for a rod average burnup of 37.3 GWd/tU. 389

Nuclear Engineering and Design 332 (2018) 383–391

400

Outer Surface Peak Cladding Hoop Stress (MPa)

Inner Surface Peak Cladding Hoop Stress (MPa)

N. Capps et al.

380 360 340 320 300 280

37.3 GWd/tU 36.1 GWd/tU 35 GWd/tU 33.9 GWd/tU 32.8 GWd/tU 31.8 GWd/tU

260 240 220 0

50

100

150

200

250

300

350

400

450

500

550

600

450

400

350

300 37.3 GWd/tU 36.1 GWd/tU 35 GWd/tU 33.9 GWd/tU 32.8 GWd/tU 31.8 GWd/tU

250

200 0

50

100

150

200

250

300

350

400

450

500

550

600

FeCrAl Coating Thickness (microns)

FeCrAl Coating Thickness (microns)

(a)

(b)

Fig. 11. R-θ BISON results comparing cladding peak hoop stress on a) inner surface and b) outer surface of the cladding as a function of coating thickness and rod average burnup. 0.06

0.05

Plastic Hoop Strain (%)

demonstration, the formation of plastic strains and differential stress increases the possibility for spallation to occur. Furthermore, with the inclusion of irradiation-induced and plastic strain hardening from the base irradiation, the FeCrAl layer would be expected to lose ductility and significantly increasing the imposed hoop stress in both the cladding and FeCrAl coating, further increasing the stress differential between the two materials. Future work would require the development and incorporation of a FeCrAl irradiation hardening model into BISON, plastic strain hardening associated with the base irradiation, a thorough ramp rate sensitivity analysis, and finally a missing pellet surface sensitivity analysis. The work presented in this article simply conducts a fundamental feasibility analysis, and detailed analysis are required in order to determine the coated cladding mechanical response to common industry issues. However, fuel performance simulations will be unable to determine if the coating material will spall or not without a damage model. Experimental data would be required to develop a spalling failure damage model.

37.3 GWd/tU 36.1 GWd/tU 35 GWd/tU 33.9 GWd/tU 32.8 GWd/tU 31.8 GWd/tU

0.04

0.03

0.02

0.01

0.00

0

100

200

300

400

500

600

FeCrAl Coating Thickness (microns)

Fig. 12. R-θ outer surface plastic hoop strain trend as a function of coating thickness and rod average burnup.

5. Conclusions and future work

analysis. BISON does calculate plastic strain during the local effects analysis, and its contributes to plastic strain hardening: however, the FeCrAl yield stress will be lower than expected. With the FeCrAl yield stress being lower than expected, the global cladding stiffness will be lower as well. Future work should include the development and implementation of a FeCrAl irradiation hardening model as well as a mechanism to include plastic strain hardening accumulated during the base irradiation. It remains unclear how the differential tensile hoop stress, illustrated in Fig. 11, will impact the FeCrAl coating’s potential for spalling. In order to understand spalling potential, a damage model based on experimental data is required. However, by turning the FeCrAl plasticity model on, the inner surface hoop stress increases while the outer surface decreases, showing an inverse dependence between the two. However, the additional plastic strain reflected by the larger stress difference with plasticity invoked further increases the potential for separation between the coating and parent cladding material. Further evidence, in Fig. 13, illustrates the hoop stress distribution in the cladding and coating for an elastic-plastic FeCrAl layer Fig. 13a). Fig. 13c shows the FeCrAl plastic strain associated with the hoop stress contour plot in Fig. 13b. The images in Fig. 13 show the importance to understanding the coating-cladding interface as well as accounting for all material properties in the coating. While this problem is a

This paper has presented the results of PCI analyses on mechanical responses of both Zr-4 cladding with an exterior layer of FeCrAl coating and FeCrAl cladding and compared their responses to that of traditional Zr-4 cladding materials. Data extracted from an EPRI report contained PWR pre-irradiated rod characteristic properties, Diablo Canyon Unit 2 Cycle 8 steady state operations, and Cycle 9 reactor startup based on Westinghouse recommendations (EPRI Report, 1999). Subsequently, BISON was used to evaluate a variety of fuel rod average heat rates typically observed in commercial PWRs which established multiple rod conditions. The analysis objectives were to: 1) compare fuel rod conditions following an initial base irradiation cycle for a variety of rod average burnups, 2) evaluate cladding hoop stress conditions at the peak power during reactor startup in order to assess the impacts on PCI response, and 3) determine the effect of FeCrAl plasticity on local PCI response. Using the BISON fuel performance code 2-D R-Z modeling capabilities showed the fuel rod response to be highly dependent on the cladding mechanical properties. The analysis demonstrated as the FeCrAl coating thickness is increased, the cladding exhibits a stiffer mechanical response. Fuel temperatures were increased early on in life, as a result of the stiffer mechanical response reducing cladding creep down, increasing fuel pellet swelling as well as increasing the fission gas released. This relationship was exhibited for all rod average 390

Nuclear Engineering and Design 332 (2018) 383–391

N. Capps et al.

Fig. 13. R-θ stress contour comparisons between 100 µm FeCrAl coated clad assuming a) with FeCrAl plasticity; b) shows the calculated plastic hoop strain corresponding to figure a).

burnups analyzed. Local effects, e.g. BISON2-D R-θ modeling, results demonstrated the effects of coating thickness, rod average burnup, elastic, and elastic-plastic material properties on both inner and outer surface cladding hoop stresses. As the FeCrAl coating increases in thickness, the inner surface hoop stress decreased, however, the relationship observed for the cladding outer surface is inversely related. As the rod average burnup decreases the hoop stress is increased on both the inner and outer surfaces. This is attributed to a reduction in the pellet-cladding gap as the rod average burnup decreases. The comparison between the elastic and plastic model results found that the inner surface hoop stress increases as plastic deformation occurs. Plastic deformation in general has a strain hardening effect, making the FeCrAl stiffer and less likely to yield to the expanding fuel pellet. Lastly, higher pellet temperatures result in more fission gas being released. Some fission gas products have been linked to SCC, and an increase in released fission gas could increase the potential for PCI-SCC to occur. This assessment has analyzed the mechanical response of FeCrAl coated Zr-4 and FeCrAl cladding types under PCI conditions and compared those responses to traditional Zr-4 cladding. Observations have shown cladding inner surface hoop stresses are reduced as a FeCrAl coating is applied, however, outer surface clad stresses are dramatically increased. This leads to concerns on coating delamination. Furthermore, all local effects simulations did not account for irradiation hardening or plastic strain hardening accumulated from the base irradiation, both of which could potentially affect the FeCrAl yield stress, change the cladding mechanical state. Future work should include the development and implementation of a FeCrAl irradiation hardening model and incorporate a mechanism to account for an initial plastic strain hardening in the local effects simulations. While this assessment has evaluated and discussed the mechanical performance of FeCrAl coated Zr-4 under typical PCI conditions, further investigations on startup power ramp rates and the development of a spalling failure model are necessary to better understand the coated cladding mechanical response.

Capps, N., Montgomery, R., Sunderland, D., Pytel, M., Wirth, B.D., 2016. Evaluation of missing pellet surface geometry on cladding stress distribution and magnitude. Nucl. Eng. Des. 305, 51. Carmack, J., Goldner, F., Bragg-Sitton, S.M., Snead, L., 2013. Overview of the U.S. DOE accident tolerant fuel development program. In: Proceedings of Top Fuel 2013, September 15–19, 2013, Charlotte, NC, USA. Cox, B., 1990. Pellet cladding interaction (PCI) failures of zirconium alloy fuel cladding – a review. J. Nucl. Mater. 172, 249–292. EPRI Report TR-100408, 1992. Corrosion of Zircaloy-Clad Fuel Rods in HighTemperature PWRs: Measurement of Waterside Corrosion in North Anna Unit 1’. EPRI Report TR-112140-V2, 1999. Effect of Startup Ramp Rate on Pellet-Cladding Interaction of PWR Fuel Rods. Garzarolli, F., von Jan, R., Stehle, H., 1979. The main causes of fuel element failure in water-cooled power reactors. At. Energy Rev. 17, 1. Gaston, D., Newman, C., Hansen, G., Lebrun-Grandie, D., 2009. MOOSE: parallel computational framework for coupled systems of nonlinear equations. Nucl. Eng. Des. 239, 1768–1778 209. Groeschel, F., Bart, G., Montgomery, R., Yagnik, S.K., 2002. Failure root cause of a PCI suspect liner fuel rod. In: IAEA Technical Meeting on Fuel Failure, Bratislava, Slovakia, 17–21 June 2002. Kiran Kumar, N.A.P., Stevens, J.N., Savela, M., Mays, B., Strumpell, J., 2016. AREVA enhanced accident tolerant fuel program – current results and future plans. In: Proceedings of Top Fuel 2016, September 2016, Boise, ID, USA. Lyons, M.F., Coplin, D.H., Jones, C.G., 1963. GE Quarterly Progress Reports, GEAP-377110 to -12 (1963/64). Montgomery, R.O. et al., Peregrine: Validation and Benchmark Evaluation of Integrated Fuel Performance Modeling Using Test Reactor Data and Falcon L1 CASL.P7.02, June, 2013. Nesbit, S., Kennard, M., Yagnik, S., 2009. Use of core analyses in assessments of fuel failure risk due to pellet-cladding interaction. In: Proc. ANS 2009, Topical Meeting ANFM 2009. pp. 1–20. Rashid, Y.R., Zangari, A.J., Lin, C.L., 1988. Modeling of PCI under steady state and transient operating conditions. In: IAEA Technical Committee Meeting on Water Reactor Fuel Element Computer Modeling in Steady State, Transient and Accident Conditions, IAEA-659/IWGFPT/32 Preston England. pp. 91–101, September 18–22, 1988. Rebak, R.B., Kim, Y.J., Gynnerstedt, J., Terrani, K.A., Stachowski, R.E., 2016. Fabrication of FeCrAl cladding for accident tolerant fuel. In: Proceedings of Top Fuel 2016, September 2016, Boise, ID, USA. Roberts, J.T.A., Gelhaus, F.E., 1979. Zircaloy performance in light water reactors. In: Zirconium in the Nuclear Industry (Fourth Conference), ASTM STP 681, ASTM, 1979, pp. 19–39. Seiler, P., Bäker, M., Rösler, J., 2011. Variation of creep properties and interfacial roughness in thermal barrier coating systems. Adv. Ceram. Coat. Mater. Extreme Environ. 32, 129–136. Stachowski, R.E., Rebak, R.B., Gassmann, W.P., Williams, J., 2016. Progress of GE development of accident tolerant fuel FeCrAl cladding. In: Proceedings of Top Fuel 2016, September 2016, Boise, ID, USA. Sunderland, D., Lyon, W., Montgomery, R., 1999. PWR Restart Ramp Rate Restrictions, EPRI, Palo Alto, CA, Electicité de France-SEPTEN, Cedex, France, and Korea Electric Power Corporation, Munji-Dong, Yusung-Gu, Taejon, Korea: 1999. TR-112140-V1. Terrani, K.A., Karlsen, T.M., Yamamoto, Y., 2016. Input correlations for irradiation creep of FeCrAl and SiC based on in–pile Halden test results. Technical Report ORNL/TM2016/191, ORNL, May 2016. Williamson, R.L., Capps, N.A., Liu, W., Rashid, Y.R., Wirth, B.D., 2016. Multi-Dimensional Simulations of LWR Fuel Behavior in the BISON Fuel Performance Code. JOM (ISSN 1047-4838).

Acknowledgement This material is based upon work supported by the Department of Energy under Award Number DENE0008416. References Capps, N., Montgomery, R., Sanders, D., Wirth, B.D., 2016. PCI Analysis of a commercial PWR using BISON fuel performance code. In: Proceedings of Top Fuel 2016, September 2016, Boise, ID, USA.

391