Using Falcon to Develop RIA Pellet-Cladding Mechanical Interaction (PCMI) Failure Criteria

Authors:
Introduction

The goal to achieve higher fuel rod burnup levels has produced considerable interest in the transient response of high burnup nuclear fuel. Several experimental programs have been initiated to generate data on the behavior of high burnup fuel under transient conditions representative of Reactivity Initiated Accidents (RIAs). A RIA is an important postulated accident for the design of Light Water Reactors (LWRs). It is considered the bounding accident for uncontrolled reactivity insertions.

The initial results from RIA-simulation tests on fuel rod segments with burnup levels above 50 GWd/tU, namely CABRI REP Na-1 (conducted in 1993) and NSRR HBO-1 (conducted in 1994), raised concerns that the licensing criteria defined in the Standard Review Plan (NUREG-0800) may be inappropriate beyond a certain level of burnup. Figure 1 is an example of a typical high burnup fuel cladding showing the oxidized and hydrided cladding of higher burnup fuel rods. Figure 2 shows the typical radial crack path in oxidized and hydrided cladding, subjected to RIA simulation tests. As a consequence of these findings, EPRI with the assistance of the Structural Integrity’s Nuclear Fuel Technology Division (formally ANATECH) and other nuclear industry members conducted an extensive review and assessment of the observed behavior of high burnup fuel under RIA conditions. The objective was to conduct a detailed analysis of the data obtained from RIA-simulation experiments and to evaluate the applicability of the data to commercial LWR fuel behavior during a Rod Ejection Accident (REA) or Control Rod Drop Accident (CRDA). The assessment included a review of the fuel segments used in the tests, the test procedures, in-pile instrumentation measurements, post-test examination results, and a detailed analytical evaluation of several key RIA-simulation tests.

This postulated accident results from an inadvertent insertion of reactivity due to the ejection of a control rod assembly in a Pressurized Water Reactor (PWR) or the drop of a control blade in a Boiling Water Reactor (BWR). In the unlikely event that sufficient reactivity is inserted into the reactor core by the ejected/dropped control rod, prompt energy deposition into the fuel can occur, which when sufficiently high can lead to fuel rod failure or, at large energy deposition levels, expulsion of UO2 fragments or molten UO2 material from the fuel rod. Requirements to mitigate the consequences of an RIA are specified within the regulations used to license and operate LWRs.

The schematic in Figure 3 highlights the relationships between the power pulse, the energy deposition and the radial average fuel enthalpy. The energy deposition represents the integration of the power-time curve and reaches the total energy deposited once the power returns to zero. The radial average fuel enthalpy is calculated based on the UO2 specific heat and the radial temperature profile. A maximum is reached near the late part of the power pulse as heat conduction effects begin to dominate. The relative response of these different parameters depends on the pulse width defined by the full-width half maximum (FWHM) of the power pulse.

It was also found that the RIA-simulation test conditions are not representative of those expected during a postulated in-reactor REA or CRDA. The experiments were conducted either in room-temperature, atmospheric-pressure water or in hot sodium coolant. The pulses were considerably more rapid (sharper and narrower) than anticipated LWR power pulses calculated using 3-D spatial kinetics methods. Additionally, in many cases, the conditions under which the test rods were base-irradiated produced cladding corrosion and hydriding features that were not representative of commercial LWR fuel. Therefore, analytical evaluations and separate effects data were required to understand the key mechanisms operative in RIA-simulation tests and to translate the experimental results to LWR conditions and different cladding materials. The key finding of the assessment was that loss of cladding ductility, due to increased localized hydrogen content, was the major cause of failure for high burnup test rods during the RIA-simulation tests.

The results from RIA-simulation experiments performed to evaluate the transient behavior of high burnup fuel have shown that the original fuel rod acceptance criteria defined in NUREG-0800 may be insufficient to insure compliance with safety requirements beyond a certain level of burnup for some postulated reactivity-initiated transients. In response to these observations, Section 4.2 of the Standard Review Plan [NUREG-0800] was amended by the NRC to include interim guidance for RIA events. Appendix B, Interim Acceptance Criteria and Guidance for the RIA was included in Revision 3 of the SRP.

As a logical next step in the process, the Regulatory Technical Advisory Committee (Reg-TAC) of the EPRI-sponsored Fuel Reliability Program, with the Nuclear Fuel Technology Division’s assistance, developed a strategy to resolve the RIA licensing issues raised by the RIA-simulation experiments and the publication of the interim criteria in NUREG-0800, Section 4.2, Appendix B, Revision 3.

The approach employed to develop the suggested revised licensing criteria combined elements of experimental data and analytical evaluations to establish a fundamental understanding of fuel behavior during RIA events. This approach was comprised of three major components:

  • 1. Establishing the transient behavior of intermediate and high burnup fuel rods using well-characterized RIA simulation tests. The RIA-simulation experiments in the previous evaluation, and the more recent tests on rods with burnup levels ranging from 45-77 GWd/tU in CABRI and NSRR, provided a database of in-pile observations and post-test examinations that can be used to evaluate the phenomena and mechanisms that influence the transient performance of the fuel and cladding under these conditions.
  • 2. Defining the cladding mechanical properties using data from separate effects tests. The database of Zircaloy cladding mechanical properties furnished insights into the influence of irradiation damage, hydrogen content and distribution, and temperature on the capability of the cladding to accommodate the pellet loading during an RIA event.
  • 3. Benchmarking the RIA analysis capabilities with Falcon using experimental data from the database of RIA-simulation tests. Performing fuel rod analyses of the RIA experiments provides a means to validate the predictive capabilities of the program and provides insights into the mechanisms that influence the pellet and cladding transient performance.

Detailed fuel behavior analyses were performed by Structural Integrity for key RIA-simulation experiments using the EPRI fuel behavior code Falcon. These analyses were based on the CSED/SED approach presented in Reference 1. Starting from room temperature and atmospheric pressure, reactivity insertion leads to a very short power pulse in the test rod. The energy deposited in the fuel causes the fuel to expand and close the gap very quickly. With further energy deposition and fuel expansion, the cladding is strained and at the peak power, the combination of high strain rate and relatively lower cladding temperature gives rise to the strong constraint on the cladding.

After that, though the fuel is still expanding, the cladding yield stress starts to decrease; as a result, peak hoop stress is achieved. Since the mechanical load is driven by fuel thermal expansion, which correlates to the total energy deposited in the fuel, both the peak hoop stress and the maximum Strain Energy Density (SED) are reached in the pulse phase. Cladding mechanical properties are then used to develop the Critical Strain Energy Density (CSED) required to initiate material failure based on mechanical property tests conducted with irradiated Zircaloy cladding. The CSED is represented as a function of the material condition, temperature, and loading state. An increase in the potential for cladding failure is assumed to occur at the point where the Falcon calculated SED exceeds the CSED for the given cladding condition defined by temperature and hydrogen content. These analyses were then used to calculate scaling factors to adjust experimental data generated at non-prototypic cold reactor coolant conditions and pressures to hot power conditions.

For the PWR hot zero power conditions, the scaled test results are presented in Figure 4 and compared to the interim NRC criterion. These data points were then fit to generate a failure enthalpy limit as a function of hydrogen. However, it is possible for BWR CRDA to occur when the coolant temperature is near ambient conditions. Nevertheless, operating history shows that most of the time during startup criticality occurs at 50°C or higher. Although an RIA event can occur prior to core wide criticality, such an event early in core startup will not result in significant energy deposition because the reactivity increase from such an event needs to first overcome the reactivity difference to core wide criticality before any energy deposition can take place (control rods are withdrawn in banks and core wide critically is reached gradually).

Also, considering that the neutronics associated with CRD events results in 0pulse widths on the order of 20 ms or more, the mid-wall cladding temperature would experience an increase of approximately 50°C. Therefore, the cladding temperature for BWR events will be on the order of 80°C or higher. Mechanical property tests show that the ductile/brittle transition begins at approximately 60°C to 70°C and is dependent on the loading rate. For temperatures >85 °C and pulse widths > 15 ms, the cladding material is shown to be ductile and therefore failure by PCMI is not likely below injected enthalpies of 150 cal/g. For ambient temperature conditions, the SED/CSED approach with Falcon was used to determine the enthalpy increase associated with scaling the room temperature NSRR data to pulse widths of 5ms and > 15ms. The increased enthalpies accounting for the loading rate effect are presented in Figure 5.

The overall fuel rod failure threshold was obtained by combining the high temperature failure threshold of 150 cal/gm with the fuel enthalpy required to produce cladding failure by PCMI as determined by the analytical evaluation. The decrease in the failure threshold is caused by two factors, the increase in PCMI loading due to gap closure effects at higher fuel rod burnups and by the decrease in cladding ductility with hydrogen accumulation. Pulse width and temperature were found to be the two factors that improve cladding ductility and therefore improving PCMI loading resistance. These results demonstrate the capability of the methodology to conservatively model the complex thermal and mechanical behavior of high burnup fuel during rapid energy depositions corresponding to a RIA event. Applying fuel cladding mechanical properties at commercial reactor conditions significantly increased the ability of the fuel to absorb energy without failure during a RIA. These new RIA PCMI acceptance criteria have been proposed for the final version of the interim RIA failure criteria for PCMI processes still under consideration by the NRC staff.

Acknowledgments

This work was conducted under the auspices of and with the support of the Regulatory Technical Advisory Committee of the EPRI Fuel Reliability Program.

References

[1] Fuel Reliability Program: Proposed Reactivity Insertion Accident (RIA) Acceptance Criteria, Revision 1. EPRI, Palo Alto, CA: 2015. 3002005540.

Contact Form