Mesh-free numerical methods provide flexible discretisations for complex geometries; however, classical meshless discrete differential operators typically trade low computational cost for limited accuracy or high accuracy for substantial per-stencil computation. We introduce a parametrised framework for learning mesh-free discrete differential operators using a graph neural network trained via polynomial moment constraints derived from truncated Taylor expansions. The model maps local stencils relative positions directly to discrete operator weights. The current work demonstrates that neural networks can learn classical polynomial consistency while retaining robustness to irregular neighbourhood geometry. The learned operators depend only on local geometry, are resolution-agnostic, and can be reused across particle configurations and governing equations. We evaluate the framework using standard numerical analysis diagnostics, showing improved accuracy over Smoothed Particle Hydrodynamics, and a favourable accuracy-cost trade-off relative to a representative high-order consistent mesh-free method in the moderate-accuracy regime. Applicability is demonstrated by solving the weakly compressible Navier-Stokes equations using the learned operators.
In the UK, many fossil-fuel power stations are approaching retirement, yet these sites retain valuable brownfield assets, notably developed industrial land and high-capacity grid connections, and can be repurposed for low-carbon generation. Repurposing such sites for low-carbon generation could therefore offer three linked advantages: lower upfront capital requirements through reuse of grid connections and developed industrial land; shorter project delivery times by reducing permitting needs and avoiding lengthy grid-connection queues; and higher system value through the provision of dispatchable flexibility and improved asset utilisation. This study examines repurposing end-of-life gas power stations into modular nuclear plants and further enabling hydrogen integration via high-temperature electrolysis, where the electricity-heat split is governed by endogenous reactor choice. A repurposing-oriented optimisation framework is proposed to embed inherited brownfield constraints, particularly land availability and grid-export limits, while co-optimising technology selection with hourly unit commitment and economic dispatch. Under the selected representative-week simulations, results show that the repurposed modular nuclear can deliver system-level flexibility comparable to that of gas turbines while meeting brownfield constraints. Nuclear‑hydrogen integration further increases reactor utilisation by using co-located electrolysis as a flexible local demand for nuclear-derived electricity and heat, eliminating renewable curtailment and substantially reducing plant-level levelised cost of electricity (LCOE) relative to power-only repurposing under these simulations. Among the electrolyser options, high-temperature electrolysis coupling gives the lowest plant-level levelised cost of hydrogen (LCOH), while producing the largest hydrogen output with the lowest plant-level upfront investment. This advantage arises from the ability of high-temperature electrolysis to use both electricity and cost-effective nuclear heat, reducing reliance on electricity-priced input energy. Sensitivity analysis identifies nuclear and high-temperature electrolyser capital costs as the main upward cost drivers, while stack lifetime remains the key durability-related lever. Overall, the findings indicate that brownfield nuclear‑hydrogen integration can provide a valuable pathway for converting retiring thermal power station sites into flexible low-carbon power and hydrogen production assets.
Abstract Critical infrastructure systems (CISs) are increasingly exposed to natural hazards whose impacts are amplified by infrastructure interdependence, urban complexity, and the growing prevalence of compound and cascading events. Existing research provides valuable insights into individual hazards, infrastructure components, modeling methods, and resilience metrics, but these insights remain difficult to translate into preliminary decision support because they are organized across different hazard types, sectors, analytical scales, and decision contexts. This article presents a structured state-of-the-art review of natural hazards, CIS impacts, and resilience enhancement. A two-pathway review strategy was adopted. A Web of Science (WoS) core search was used to characterize the bibliometric landscape and identify WoS-sourced core synthesis studies, while a targeted anchor-search pathway was used to verify foundational concepts, hazard mechanisms, methodological approaches, resilience frameworks, and authoritative institutional sources. The review clarifies the conceptual distinction among dependencies, interdependencies, compound hazards, cascading failures, service disruption, and resilience; synthesizes cross-hazard infrastructure impact mechanisms; compares major method families for hazard-impact and resilience assessment; and develops a synthesis-derived rapid resilience assessment framework for preliminary screening and prioritization under constrained data and time conditions. The synthesis indicates that future CIS resilience assessment would benefit from moving beyond isolated asset-damage analysis toward integrated workflows that link hazard loading, component vulnerability, network functionality, service consequences, recovery capacity, uncertainty, and feasible intervention choices. The article contributes a more transparent evidence-source structure, an evidence-to-decision bridge, and a bounded decision-oriented framework for organizing future research and practical resilience screening.
Armed conflict can disrupt food systems simultaneously through reduced production, damaged infrastructure, insecure transport corridors, export restrictions, fertilizer scarcity, energy-price inflation, and diminished purchasing power. This chapter examines the organizational consequences of these interacting shocks through grain and fertilizer supply chains affected by the Russia–Ukraine war. It develops a source–corridor–input vulnerability framework connecting origin concentration, route concentration, input criticality, substitution time, and buffering capacity. Scenario-based analysis shows that alternative maritime, river, rail, and road corridors can preserve physical availability but frequently increase handling, freight, insurance, storage, and financing costs. Fertilizer and energy shocks can also outlast the immediate interruption of grain exports by reducing subsequent harvests. Resilience therefore requires simultaneous protection of supply continuity, commercial affordability, agricultural production, and access for vulnerable consumers.
Computational models of memory have achieved considerable success by formalizing how traces are encoded, cued, and retrieved. However, the role that individual linguistic experience plays in shaping representational structure has been relatively understudied. Where the issue has been examined, most models assume that a common population-level semantic space suffices, effectively treating individual variation in language experience as noise rather than signal. We test the theoretical adequacy of this assumption. In a large-scale experiment (N = 478), participants completed a cued recall task in which identical target words were paired with cues drawn from eight different genre-specific semantic spaces (mystery, fantasy, horror, literary fiction, romance, science fiction, thriller, and non-fiction). Participants reported their reading habits across the eight genres, and personalized semantic representations were constructed by weighting genre-specific corpora proportionally to each participant's reading profile. Two complementary analyses were conducted. In a model-free analysis, cosine similarity computed within each participant's personalized semantic space predicted recall and omission outcomes more reliably than similarity derived from generic semantic spaces. In a computational analysis, substituting personalized representations into the embedded Computational Framework of Memory improved cue-target-level predictions over generic alternatives, accounting for approximately 4 percentage points more explained variance. These findings provide proof of concept that variation in individual linguistic experience shapes semantic structure in ways that are both behaviourally detectable and computationally tractable. More broadly, they suggest that the predictive limits of memory models may reside not only in their retrieval mechanisms but also in the fidelity of the representations they assume.
