CASL

Science and Technology

Validation of Peregrine with Test Reactor Data

At the end of September, Pellet-Cladding Interaction (PCI) Challenge Problem Integrator Robert Montgomery reported that good progress has been made in demonstrating the Peregrine LWR fuel performance modeling software.  The Peregrine fuel performance analysis computer program is being developed to provide a single rod 3-dimensional fuel performance modeling capability to assess safety margins and the impact of plant operation and fuel rod design on the thermo-mechanical behavior of nuclear fuel.

Peregrine is built using the Multiphysics Object Oriented Simulation Environment (MOOSE) computational framework architecture developed at Idaho National Laboratory, and uses the finite element method for geometric representation and a Jacobian-free, Newton-Krylov scheme to solve systems of partial differential equations. Peregrine also leverages key capabilities built into the Bison general purpose fuel modeling system. The ability to employ massively parallel computational capabilities is a key advantage to using Peregrine to study nuclear fuel performance, especially for multi-dimensional, multi-physics phenomena. 

Many of the performance models used in Peregrine originated with EPRI’s fuel performance code, Falcon.  Falcon is an enhanced and integrated derivative of ESCORE and FREY, two historic EPRI codes developed in the 1980s with final release in 1990 and 1991 respectively. Enhancements continued from 1996 to 2004 when FALCON MOD01 Beta was released as the state‑of-the-art LWR fuel performance code validated to high burn-up, capable of analyzing both steady state and transient fuel behavior.  Falcon, now a brand name, is a result of further enhancements to make fuel performance analysis more accusable to nuclear utility fuel managers and reactor engineers.

Pellet-cladding mechanical interaction, or PCMI, refers to the contact between the fuel cladding and fuel pellet that is a normal part of fuel rod performance, typically occurring towards the end of the rod’s first operational cycle.  In the early 2000’s several leaking rods were observed in LWRs and the root cause was found to be a combination of mechanical and chemical interactions between fuel pellets and fuel rod cladding, commonly called PCI. A necessary condition for these failures is contact between the pellet and cladding and a generated stress in the cladding caused by a power change.

PCI failures are often referred to as classical and non-classical type failures.  Classical PCI failure refers to stress corrosion cracking (SCC) induced cladding fractures initiated during a power increase.  Non-classical PCI failures can occur under less severe operating conditions where other factors, such as a missing pellet surface, overstress the cladding.  PCI failures may occur in both PWRs and BWRs.  The failure mechanism is more prevalent in BWRs because control rod movements are more frequent.  In PWRs, reactor power is typically controlled through the addition of soluble boron in the coolant, and consequently PCI failures are less frequent.  However, during reactor power increases, and specifically during a class II transient (anticipated operational occurrence), PCI failures may occur in a PWR.

Thus, to reliably predict PCI failures, the code must be capable of describing the thermal, mechanical, and chemical properties and constitutive relationships of both pellet and cladding as functions of temperature, chemistry, burnup, fission density, fast flux, fast fluence, and many other state variables. 

The non-linear fuel rod behavior resulting from the interactions of these properties require key assumptions concerning material dependencies, numerical formulations, and geometric representation to satisfy runtime, numerical convergence, and assumption limitations, giving rise to simplifications in the approaches to de-emphasize one phenomenon over others. Consequently, fuel performance modeling software requires extensive verification, calibration, and validation exercises to demonstrate their ability to accurately represent nuclear fuel behavior.

The current validation work is based on fourteen fuel rods irradiated in a variety of commercial and test reactors to assess the fuel temperature, cladding deformation, and fission gas release modeling capabilities.  These 14 fuel rods experienced irradiation conditions representative of beginning of life all the way to high burnup.  Because of the nature of the irradiation programs, these fuel rods were modeled using axis-symmetric geometric representations of the cylindrical fuel pellet column and cladding as shown in Figure 4.  A comparison of Peregrine to fuel centerline temperatures from more than 550 measurements is shown in Figure 5, along with a comparison to the industry fuel performance modeling software package, Falcon.

The validation activities provide confidence in the development of Peregrine to date, but also highlight the challenges in accurately modeling the complex thermal and mechanical behavior inherent in nuclear fuel performance. A number of improvements in the material and behavior models have been identified by the development team, and future development activities will focus on enhancing and implementing these models into Peregrine. In particular, improvements to the representation of pellet cracking and relocation, fission gas retention and release, gap thermal conductance, pellet-clad mechanical contact (including improved advanced creep and growth models) and cladding oxidation and hydride formation and growth are needed.  These advancements will expand the fidelity of Peregrine and provide the ability to accurately model three-dimensional aspects of fuel performance, such as pellet-clad interaction with missing pellet surfaces. In some cases, advanced models for these topics are being developed within the Materials Performance and Optimization (MPO) focus area and will be implemented in Peregrine to enhance code capability.

For more information, see CASL-I-2013-0122-000 and CASL-I-2013-0165-000.