-
Predictions and Uncertainty Estimates of Reactor Pressure Vessel Steel Embrittlement Using Machine Learning
Authors:
Ryan Jacobs,
Takuya Yamamoto,
G. Robert Odette,
Dane Morgan
Abstract:
An essential aspect of extending safe operation of the active nuclear reactors is understanding and predicting the embrittlement that occurs in the steels that make up the Reactor pressure vessel (RPV). In this work we integrate state of the art machine learning methods using ensembles of neural networks with unprecedented data collection and integration to develop a new model for RPV steel embrit…
▽ More
An essential aspect of extending safe operation of the active nuclear reactors is understanding and predicting the embrittlement that occurs in the steels that make up the Reactor pressure vessel (RPV). In this work we integrate state of the art machine learning methods using ensembles of neural networks with unprecedented data collection and integration to develop a new model for RPV steel embrittlement. The new model has multiple improvements over previous machine learning and hand-tuned efforts, including greater accuracy (e.g., at high-fluence relevant for extending the life of present reactors), wider domain of applicability (e.g., including a wide-range of compositions), uncertainty quantification, and online accessibility for easy use by the community. These improvements provide a model with significant new capabilities, including the ability to easily and accurately explore compositions, flux, and fluence effects on RPV steel embrittlement for the first time. Furthermore, our detailed comparisons show our approach improves on the leading American Society for Testing and Materials (ASTM) E900-15 standard model for RPV embrittlement on every metric we assessed, demonstrating the efficacy of machine learning approaches for this type of highly demanding materials property prediction.
△ Less
Submitted 5 September, 2023;
originally announced September 2023.
-
Atom probe characterisation of segregation driven Cu and Mn-Ni-Si co-precipitation in neutron irradiated T91 tempered-martensitic steel
Authors:
T. P. Davis,
M. A. Auger,
N. Almirall,
P. Hosemann,
G. R. Odette,
P. A. J. Bagot,
M. P. Moody,
D. E. J. Armstrong
Abstract:
The T91 grade and similar 9Cr tempered-martensitic steels (also known as ferritic-martensitic) are leading candidate structural alloys for fast fission nuclear and fusion power reactors. At low temperatures (300 to 400 $^\circ$C) neutron irradiation hardens and embrittles these steels, therefore it is important to investigate the origin of this mode of life limiting property degradation. T91 steel…
▽ More
The T91 grade and similar 9Cr tempered-martensitic steels (also known as ferritic-martensitic) are leading candidate structural alloys for fast fission nuclear and fusion power reactors. At low temperatures (300 to 400 $^\circ$C) neutron irradiation hardens and embrittles these steels, therefore it is important to investigate the origin of this mode of life limiting property degradation. T91 steel specimens were separately neutron irradiated to 2.14 dpa at 327 $^\circ$C and 8.82 dpa at 377 $^\circ$C in the Idaho National Laboratory Advanced Test Reactor. Atom probe tomography was used to investigate the segregation driven formation of Mn-Ni-Si-rich (MNSPs) and Cu-rich (CRP) co-precipitates. The precipitates increase in size and, slightly, in volume fraction at the higher irradiation temperature and dose, while their corresponding compositions were very similar, falling near the Si(Mn,Ni) phase field in the Mn-Ni-Si projection of the Fe-based quaternary phase diagram. While the structure of the precipitates has not been characterized, this composition range is distinctly different than that of the typically cited G-phase. The precipitates are composed of CRP with MNSP appendages. Such features are often observed in neutron irradiated reactor pressure vessel (RPV) steels. However, the Si, Ni, Mn, P and Cu solutes concentrations are lower in the T91 than in typical RPV steels. Thus, in T91 precipitation primarily takes place in solute segregated regions of line and loop dislocations. These results are consistent with the model for radiation induced segregation driven precipitation of MNSPs proposed by Ke et al. Cr-rich alpha prime ($α$') phase formation was not observed.
△ Less
Submitted 21 July, 2020;
originally announced July 2020.
-
Flux effects in precipitation under irradiation - Simulation of Fe-Cr alloys
Authors:
Jia-Hong Ke,
Elaina R. Reese,
Emmanuelle A. Marquis,
G. Robert Odette,
Dane Morgan
Abstract:
Radiation-enhanced precipitation of Cr-rich α' in irradiated Fe-Cr alloys, which results in hardening and embrittlement, depends on the irradiating particle and the displacement per atom (dpa) rate. Here, we utilize a Cahn-Hilliard phase-field based approach, that includes simple models for nucleation, irradiating particle and rate dependent radiation-enhanced diffusion and cascade mixing to simul…
▽ More
Radiation-enhanced precipitation of Cr-rich α' in irradiated Fe-Cr alloys, which results in hardening and embrittlement, depends on the irradiating particle and the displacement per atom (dpa) rate. Here, we utilize a Cahn-Hilliard phase-field based approach, that includes simple models for nucleation, irradiating particle and rate dependent radiation-enhanced diffusion and cascade mixing to simulate α' evolution under neutrons, heavy ions, and electron irradiations. Different irradiating particles manifest very different cascade mixing efficiencies. The model was calibrated using neutron data. For cascade inducing neutron/heavy-ion dpa rates at 300 °C between 10^-^8 and 10^-^6 dpa/s the model predicts approximately constant number density, decreasing radius, decreasing α' Cr composition, and lower α' volume fraction. The model then predicts a dramatic transition to no α' formation above approximately 10^-^5 dpa/s, while electron irradiation, with weak mixing, had little effect at dpa rates up to 10^-^3 dpa/s. These model predictions are consistent with experiments. We explain the results in terms of the flux dependence of the radiation enhanced diffusion, cascade mixing, and their ratio, which all vary significantly in relevant flux ranges for neutron and cascade inducing ion irradiations. These results show that both cascade mixing and radiation enhanced diffusion must be accounted for when attempting to emulate neutron-irradiation effects using accelerated ion irradiations.
△ Less
Submitted 21 April, 2019;
originally announced April 2019.
-
Thermodynamics and kinetics of core-shell versus appendage co-precipitation morphologies: An example in the Fe-Cu-Mn-Ni-Si system
Authors:
Shipeng Shu,
Peter B. Wells,
Nathan Almirall,
G. Robert Odette,
Dane D. Morgan
Abstract:
What determines precipitate morphologies in co-precipitating alloy systems? We focus on alloys of two precipitating phases, with precipitates of the fast-precipitating phase acting as heterogeneous nucleation sites for a second phase manifesting slower kinetics. We study a FeCuMnNiSi alloy using the combination of atom probe tomography and kinetic Monte Carlo simulations. It is shown that the inte…
▽ More
What determines precipitate morphologies in co-precipitating alloy systems? We focus on alloys of two precipitating phases, with precipitates of the fast-precipitating phase acting as heterogeneous nucleation sites for a second phase manifesting slower kinetics. We study a FeCuMnNiSi alloy using the combination of atom probe tomography and kinetic Monte Carlo simulations. It is shown that the interplay between interfacial and ordering energies, plus active diffusion paths, strongly affect the selection of core-shell verses appendage morphologies. Specifically, the ordering energy reduction of the MnNiSi phase heterogeneously nucleated on a pre-existing copper-rich precipitate exceeds the energy penalty of a predominantly Fe/Cu interface, leading to initial appendage, rather than core-shell, formation. Diffusion of Mn, Ni and Si around and through the Cu core towards the ordered phase results in subsequent appendage growth. We further show that in cases with higher primary precipitate interface energies and/or suppressed ordering, the core-shell morphology is favored.
△ Less
Submitted 27 July, 2018;
originally announced July 2018.
-
Cluster dynamics modeling of Mn-Ni-Si precipitates in ferritic-martensitic steel under irradiation
Authors:
Jia-Hong Ke,
Huibin Ke,
G. Robert Odette,
Dane Morgan
Abstract:
Mn-Ni-Si precipitates (MNSPs) are known to be responsible for irradiation-induced hardening and embrittlement in structural alloys used in nuclear reactors. Studies have shown that precipitation of the MNSPs in 9-Cr ferritic-martensitic (F-M) alloys, such as T91, is strongly associated with heterogeneous nucleation on dislocations, coupled with radiation-induced solute segregation to these sinks.…
▽ More
Mn-Ni-Si precipitates (MNSPs) are known to be responsible for irradiation-induced hardening and embrittlement in structural alloys used in nuclear reactors. Studies have shown that precipitation of the MNSPs in 9-Cr ferritic-martensitic (F-M) alloys, such as T91, is strongly associated with heterogeneous nucleation on dislocations, coupled with radiation-induced solute segregation to these sinks. Therefore it is important to develop advanced predictive models for Mn-Ni-Si precipitation in F-M alloys under irradiation based on an understanding of the underlying mechanisms. Here we use a cluster dynamics model, which includes multiple effects of dislocations, to study the evolution of MNSPs in a commercial F-M alloy T91. The model predictions are calibrated by data from proton irradiation experiments at 400 °C. Radiation induced solute segregation at dislocations is evaluated by a continuum model that is integrated into the cluster dynamics simulations, including the effects of dislocations as heterogeneous nucleation sites. The result shows that MNSPs in T91 are primarily irradiation-induced and, in particular, both heterogeneous nucleation and radiation-induced segregation at dislocations are necessary to rationalize the experimental observations.
△ Less
Submitted 14 November, 2017;
originally announced November 2017.
-
Thermodynamic and kinetic modeling of Mn-Ni-Si precipitates in low-Cu reactor pressure vessel steels
Authors:
Huibin Ke,
Peter Wells,
Philip D. Edmondson,
Nathan Almirall,
Leland Barnard,
G. Robert Odette,
Dane Morgan
Abstract:
Formation of large volume fractions of Mn-Ni-Si precipitates (MNSPs) causes excess irradiation embrittlement of reactor pressure vessel (RPV) steels at high, extended-life fluences. Thus, a new and unique, semi-empirical cluster dynamics model was developed to study the evolution of MNSPs in low- Cu RPV steels. The model is based on CALPHAD thermodynamics and radiation enhanced diffusion ki- netic…
▽ More
Formation of large volume fractions of Mn-Ni-Si precipitates (MNSPs) causes excess irradiation embrittlement of reactor pressure vessel (RPV) steels at high, extended-life fluences. Thus, a new and unique, semi-empirical cluster dynamics model was developed to study the evolution of MNSPs in low- Cu RPV steels. The model is based on CALPHAD thermodynamics and radiation enhanced diffusion ki- netics. The thermodynamics dictates the compositional and temperature dependence of the free energy reductions that drive precipitation. The model treats both homogeneous and heterogeneous nucleation, where the latter occurs on cascade damage, like dislocation loops. The model has only four adjustable parameters that were fit to an atom probe tomography (APT) database. The model predictions are in semi-quantitative agreement with systematic Mn, Ni and Si composition variations in alloys character- ized by APT, including a sensitivity to local tip-to-tip variations even in the same steel. The model predicts that heterogeneous nucleation plays a critical role in MNSP formation in lower alloy Ni contents. Single variable assessments of compositional effects show that Ni plays a dominant role, while even small variations in irradiation temperature can have a large effect on theMNSP evolution. Within typical RPV steel ranges, Mn and Si have smaller effects. The delayed but then rapid growth of MNSPs to large volume fractions at high fluence is well predicted by the model. For purposes of illustration, the effect of MNSPs on transition temperature shifts are presented based on well-established microstructure-prop- erty and property-property models.
△ Less
Submitted 25 July, 2017;
originally announced July 2017.
