Heavy-duty trucks are major contributors in NOx and PM2.5 emissions in the transportation sector. Many countries are seriously considering electrification as a pathway towards cleaner freight transport. The advancement of diverse technological solutions has created multiple pathways towards the decarbonisation of heavy-duty trucks, among which battery-electric heavy-duty trucks remain in the early stages of development. This paper presents a comprehensive and systematic review of their electrification from technological, economic, and social perspectives, covering battery installation approaches, electricity refuelling solutions (e.g., charging, battery swapping, wireless charging), and typical application scenarios for short- and long-haul trucking. Key issues including electricity demand, operation scheduling, total cost of ownership, and greenhouse gas emissions are discussed, followed by an identification of current research gaps in empirical data and an examination of grid integration challenges and opportunities across different technology pathways.
Chocolate is a complex colloidal system of multiple dispersed particle types, including sugar, cocoa solids and often dairy proteins. However, the use of plant proteins in such systems remains unexplored. This study investigates the behaviour of pea protein isolate (PPI), sugar particles and their blends in oil, focussing on their effects on the rheological and crystallisation properties of cocoa butter (CB). Both PPI and sugar particles aggregated in CB, increasing viscosity and forming elastic oleogel-like systems. In blended systems, PPI disrupted sugar particle-particle interactions. Greater sugar particle aggregation was observed in sunflower oil, attributed to surface-active impurities in CB acting as dispersants for hydrophilic particles. Oil type had little influence on PPI aggregation and the modified MPQ-rcp model indicated that PPI and sugar particles exhibit similar interactions in CB. PPI accelerated CB crystallisation, whereas the presence of sugar delayed crystallisation, so that mixed particle systems had similar kinetics to pure CB. Over extended storage, concentrated particle suspensions curb the solid fat content (SFC), attributed to the restricted mobility of oil due to the high dispersed volume fractions-as visualised by Raman microscopy. At low SFC, particles are active in the fat crystal network, increasing elasticity. However, at high SFC, the elasticities were comparable to that of pure CB. This demonstrates the potential of plant protein particles to provide rheology modification benefits to sugar-oil mixtures and to reduce solid fat content, providing a foundation for the development of more sustainable and nutritious confectionery systems.
Time-dependent partial differential equations (PDEs) underpin modeling of dynamical phenomena across science and engineering, but repeatedly solving them at high fidelity remains expensive. Neural operators offer reusable surrogates for such systems, yet current approaches often lose robustness under sparse or irregular temporal supervision and can be costly to train for high-resolution or long-trajectory problems. Here we introduce the Time-Attentional Neural Operator (TANO), a composite neural operator whose stacked layers each extract multiscale spatial features, model their temporal interactions with attention, and reconstruct the spatial representation. We further develop a stochastic temporal subsampling strategy in which each training step uses only a random subset of time points, improving training efficiency while retaining one-pass full-trajectory prediction over the supervised horizon. Across five canonical PDE benchmarks, TANO reduces prediction errors by 1.87–3.94× relative to existing baselines and remains robust under sparse temporal supervision or limited data. It accurately simulates wave propagation over long trajectories in heterogeneous media, trains about 4× faster than the competing approach on high-resolution problems, and can be extended to long-horizon extrapolation through a windowed rollout scheme. Our findings establish TANO as an efficient and robust neural-operator framework that makes training on complex spatiotemporal problems more computationally tractable and supports practical surrogate modeling.
Lubrication in ball-in-socket bearings, such as those in total hip replacements, is inherently transient. The load and motion change dynamically while the lubricant cavitates in regions of diverging geometry and the bearing surfaces deform elastically and plastically. Surface roughness with amplitude comparable to the film thickness further modifies the local flow, but resolving roughness directly within a transient elastohydrodynamic (EHL) simulation is computationally prohibitive, while classical flow-factor models rely on assumptions that do not readily transfer to this regime. We present a fully transient multiscale framework, based on the Heterogeneous Multiscale Methods (HMM) for EHL in conformal bearings. A smooth macroscale Reynolds model captures the bearing geometry, deformation, dynamic loading, and velocity profiles and is coupled to representative transient microscale simulations that resolve idealised roughness and return additive adjustments to macroscale flux (ΔQ) and load-bearing capacity (ΔP). By construction, the multiscale framework reduces to the underlying smooth EHL model as the roughness amplitude tends to zero. With microscale roughness included, the multiscale model agrees with a high resolution deterministic model to within 3% in minimum thickness and pressure with a significant reduction in computational cost for the cases considered. The framework is demonstrated with ASTM F3143 reciprocating hip test velocity profiles.
Last-mile delivery research increasingly focuses on business-to-consumer (B2C) automation to satisfy consumers with an ‘I want it now’ mentality, reduce firms’ carbon footprints, and improve cost efficiency. Much less attention has been given to the possibility that there might be differences across consumer segments in their preferences for technology‑enabled or more traditional, human-mediated delivery modes. Building on random utility theory, we developed a hybrid choice model with a latent class kernel and applied it to stated choice data on delivery service preferences collected from 1,007 customers across metropolitan and regional areas of Australia. Our results reveal two latent classes of consumers with distinctly different preferences. The largest group, which we label techies, value speed and low cost last-mile delivery and prefer traditional postal services, followed by aerial drones and parcel lockers, with the effect moderated by parcel value and the availability of a safe place for unattended delivery. The second group, relational, trusting people, place a stronger emphasis on trust and interpersonal aspects of service and substantially prefer traditional postal services (assuming a safe place for delivery is available) and lockers while showing a strong dislike of aerial drones. To interpret these patterns, we draw on social exchange theory and the human-to-human (H2H) relationship model to explain how relational considerations may influence delivery mode preferences. H2H interactions matter for a subset of consumers. Importantly, our broader findings highlight the continued relevance of traditional postal services in a diversifying delivery landscape. Although no direct spatial effect emerges in the class allocation model, our results suggest that residential location (metropolitan versus regional) has a structural effect by influencing consumers’ underlying latent attitudes.
