Meeting the current demand for concrete requires not only mining tons of gravel and sand, but also burning large amounts of fossil fuel resources in cement kilning. Consequently, concrete recycling is crucial to achieving a material-efficient society, especially with the application of various categories of concrete and the goal of phasing out fossil fuels. A comparative life cycle assessment (LCA) is used to assess the engineering material footprint (EMF) and the fossil fuel material footprint (FMF) in closed-loop recycling of three types of concrete: siliceous concrete, limestone concrete, and lightweight aggregate concrete. This study aims to investigate the impact of (i) concrete categories, (ii) methods to model recycling, and (iii) using renewable energy sources on the material footprint in concrete recycling. The results highlight that the concrete recycling system can reduce 99% of the EMF and 66–93% of the FMF compared with the baseline system, in which concrete waste is landfilled. All three recycling modeling approaches indicate that concrete recycling can considerably reduce EMF and FMF compared with the baseline system, primarily resulting from the displacement of virgin raw materials. Using alternative diesels is more sensitive than adopting renewable electricity in reduction of the FMF in concrete recycling. Replacing diesel with electrolysis- and coal-based synthetic diesel for concrete recycling could even increase the FMF, while using biodiesel made from rapeseed and wood-based synthetic diesel can reduce 47–51% and 84–89% of the FMF, respectively, compared to the virgin diesel-based recycling system. Finally, we discussed the multifunctionality and rebound effects of recycling, and double-counting risk in material and energy accounting.

@en

Federated learning (FL) is increasingly considered to circumvent the disclosure of private data in mobile edge computing (MEC) systems. Training with large data can enhance FL learning accuracy, which is associated with non-negligible energy use. Scheduled edge devices with small data save energy but decrease FL learning accuracy due to a reduction in energy consumption. A trade-off between the energy consumption of edge devices and the learning accuracy of FL is formulated in this proposed work. The FL-enabled twin-delayed deep deterministic policy gradient (FL-TD3) framework is proposed as a solution to the formulated problem because its state and action spaces are large in a continuous domain. This framework provides the maximum accuracy ratio of FL divided by the device’s energy consumption. A comparison of the numerical results with the state-of-the-art demonstrates that the ratio has been improved significantly.@en

The global trade of plastic waste has raised environmental concerns, especially regarding pollution in waste-importing countries. However, the overall environmental contribution remains unclear due to uncertain treatment shares between handling plastic waste abroad and domestically. Here, we conduct a life cycle assessment of global plastic waste trade in 2022 across 18 countries and six plastic waste types, alongside three “nontrade” counterfactual scenarios. By considering the required cycling rate, which balances importers’ costs and recycling revenues, we find that the trade resulted in lower environmental impacts than treating domestically with the average treatment mix. The trade scenario alone reduced climate change impact by 2.85 million tonnes of CO2 equivalent and mitigated damages to ecosystem quality, human health, and resource availability by 12 species-years, 6200 disability-adjusted life years (DALYs), and 1.4 billion United States dollars (USD in 2013), respectively. These results underscore the significance of recognizing plastic waste trade as a pivotal factor in regulating global secondary plastic production when formulating a global plastics treaty.@en

Engineering structures of concrete are generally recognized to be rarely fully water-saturated. The water transportation process in unsaturated cementitious materials offers therefore a relevant problem the understanding of which can be crucial in assessing concrete's degradation and failure mechanisms. In recent years, the number of published papers on this topic is significantly increased. This review presents the latest advancements in determining water permeability of unsaturated cement-based materials by experimental methods and numerical modelling. The effects are summarized of water-cement ratio (w/c), curing age, particle size distribution, interfacial transition zone (ITZ) and supplementary cementitious materials (SCMs) on the permeability of partially saturated cement-based materials. Next, the underlying relationship between relative water permeability and pore structure is presented and discussed. Additionally, an insight into water transport mechanism of unsaturated concrete is proposed. Finally, some evaluative conclusions are drawn that can be instrumental for setting up future studies.

@en

Federated learning (FL) has been increasingly considered to preserve data training privacy from eavesdropping attacks in mobile-edge computing-based Internet of Things (EdgeIoT). On the one hand, the learning accuracy of FL can be improved by selecting the IoT devices with large data sets for training, which gives rise to a higher energy consumption. On the other hand, the energy consumption can be reduced by selecting the IoT devices with small data sets for FL, resulting in a falling learning accuracy. In this article, we formulate a new resource allocation problem for privacy-preserving EdgeIoT to balance the learning accuracy of FL and the energy consumption of the IoT device. We propose a new FL-enabled twin-delayed deep deterministic policy gradient (FL-DLT3) framework to achieve the optimal accuracy and energy balance in a continuous domain. Furthermore, long short-term memory (LSTM) is leveraged in FL-DLT3 to predict the time-varying network state while FL-DLT3 is trained to select the IoT devices and allocate the transmit power. Numerical results demonstrate that the proposed FL-DLT3 achieves fast convergence (less than 100 iterations) while the FL accuracy-to-energy consumption ratio is improved by 51.8% compared to the existing state-of-the-art benchmark.

@en

Permeability estimates obtained by physical experiments and by simulations differ quite significantly. When biases at the experimental side would be eliminated, the expected modest gap could be bridged by surface layer modifications of the hardened cement paste, discussed herein. This involves the formation of a fractal-like nano-particle structure in the outer hydration layer, instead of the smooth-surfaced layer obtained by the vector approach. This study firstly applies a DLA method for simulating the fractal structure of nodules in a low-density 2D setting of mono-size cement particles. Herewith it is confirmed that density of the outer hydration layer will diminish almost linearly away from the inner one, inevitably leading to permeability decline. This approach yielded also a realistic range of modifications for the outer layer, i.e., a 0.1 to 0.5 μm increased thickness. Therefore, in the normal 3D set up, the smooth vector-based hydration layer is given a stepwise thickness increase in this range, whereupon the consequences for pore geometry and topology, as well as for permeability can be investigated. The outcomes of this study are laid down in a separate publication. When further structural information on nano level will become available in the near future, the simulations can be appropriately adapted. This paper specifically proposes as a practical methodology to slightly step-wise enlarge at the packing simulation stage the fresh cement grains in proportion to their respective sizes. The herein introduced standard methodology for porosimetry and permeability estimation can readily be employed for simulating these cases. This is demonstrated for cement paste with w/c = 0.5, whereby packing density is increased by about 1% per step. This can be associated with a fictitious reduction in w/c from 0.49 to 0.45. It could be demonstrated this way that a permeability decline of one order of magnitude can be realized, i.e., at least enough to bridge the modest gap with corrected experimental permeability data. So, the final solution can be expected in this range.@en

This chapter deals with the issues related to packing of particles in nature - aggregate in the first place - and in cementitious materials. Many of these topics are extensively discussed in the international literature, so will only be briefly introduced, while additionally giving a restricted number of references. Relatively new developments, such as producing the realcrete by computer, yielding compucrete, will receive more explicit attention. The way to do this properly by DEM (discrete element method) instead of by popular RSA (random sequential addition) is stressed. Modern dynamic DEM renders possible simulating aggregate of fluvial origin as well as of crushed rock. Similarity of simulating particle packing on mesolevel (aggregate) and on microlevel (binder) was the argument to sketch modern developments in the latter field, because they provide information on permeability and is thus of paramount important for sustainability. A phenomenon receiving recently attention in the literature is Brazil Nut Effect. It is part of nature and it can significantly influence the compacted particle structure of aggregate or binder alike. Some relevant information as to cementitious materials will be provided. @en

Research on cementitious materials can provide geometric data on structural topics of interest. Such structural data should be subjected to a process called “geometric averaging” to be coupled on the engineering level with properties like strength, stiffness or permeability. The prime purpose of this work is to demonstrate that the mathematics underlying the process of going from the materials structure to engineering properties is governed by simple stereology-based constants. Particularly, Cauchy concepts are referred to for expressions relating the length of a curved line in a plane or in space, alternatively of a curved surface in space, to extensions of total projections on lines, respectively planes, in random directions. The two parameters due to Cauchy, i.e., 2/π and 1/2 will be demonstrated dominantly governing the geometric averaging process. The stereological methodology to assess the geometric material parameters and the engineering application of the outcomes are also briefly touched upon in the paper.@en