Niccolò Di Eugenio

A machine‐learning framework is designed to tackle the spinodal decomposition of coherently strained alloy microstructures, with cubic anisotropy in elastic constants. A database of phase‐field simulations, leveraging a Green's function approximation of the elastic field, is constructed for randomly chosen composition fields and widely variable misfit strain, producing a continuum variation in the phase morphology from smooth to strongly anisotropic domains. A convolutional recurrent neural network is then trained to accurately predict the full time‐evolution sequence under the explicit conditioning of the known misfit parameter, at a reduced computational cost. Extensive error analysis at the pixel level and for global descriptors is used to assess the model accuracy and evaluate its generalization capability on longer timescales and larger computational domains. The model returns a one‐to‐one match of the ground‐truth simulations over the temporal range of training sequences. Moreover, it can reliably predict average behaviors for sequences several times longer the training ones, albeit losing one‐to‐one accordance. As a proof, the NN trained model is used for reconstructing the full phase diagram of the system, achieving a degree of accuracy. The proposed framework is general and can be applied beyond the specific, prototypical system here considered, enabling high‐throughput parametric studies.

Compact fusion reactors are receiving increasing interest as a promising route for accelerating the path toward commercial fusion, thanks to their reduced size and cost. However, this compactness introduces new technological challenges, including higher radiation loads on critical functional components, such as the magnet system. Neutron shielding is therefore of utmost importance to guarantee the expected lifetime of the device, and its selection must account for the harsh environment imposed by the high radiation flux. Shielding materials should be structurally stable, not melt within the operational temperature windows, and be relatively low-cost. For nuclear reactor applications, binary compounds are typically the preferred choice as they often meet these requirements, particularly in terms of availability and cost. In this work, we present a systematic Monte Carlo analysis of more than 700 binary compounds, exposed to the neutron spectrum at the most loaded position of the vacuum vessel in a simplified model of a compact fusion reactor. Shielding performances were evaluated in a toroidal geometry in terms of neutron attenuation, power deposition, and activation, leading to the identification of several promising compositions for effective neutron shielding in future fusion applications.

Compact fusion reactors require high temperature superconductors (HTS) for the production of high field magnets capable of confining the plasma and achieving enough fusion power to reach commercial interest. The generation of high energy neutrons in the plasma, together with the consequent proliferation of secondary particles, threatens the long term performance of the HTS materials due to the heat load and radiation damage. For this reason it is crucial to develop tools for the comprehensive simulation of radiation environment and effects throughout the reactor structure. Here we present Monte Carlo simulations of an ARC-like reactor yielding the neutron and secondary particles spectra, the power deposition, and the materials damage mapped over the whole system, focusing on the effects on the toroidal field and poloidal field HTS coils. We find that the peak of the heat deposition in the magnets is located near the divertor, and that a strongly non-homogeneous distribution of damage urges the geometrical optimization of neutron shields.