Search

The Kerguelen Plateau, southern Indian Ocean, which rises up 2000 m above the surrounding seafloor, forms an obstacle for the flow of the Antarctic Circumpolar Current (ACC) and Antarctic Bottomwater (AABW). The ACC is strongly deviated in its flow towards the north. A branch of the AABW flows northwards along the eastern flank of the plateau and in its path is steered by several basement highs and William’s Ridge. Seismic data collected during RV Sonne cruise SO272 image sediment drifts shaped in the Labuan Basin, which document onset and variabilities in pathway and intensity of this AABW branch in relation to the development of the Antarctic ice sheet and tectonic processes, e.g., the opening of the Tasman gateway. The eastern flank of the Kerguelen further shows strong erosion of the post-mid Eocene sequences. In places, the Paleocene/early Eocene sequences are also affected by thinning and erosion. A moat can be observed along the Kerguelen Plateau flank indicating the flow path of the north setting AAWB branch. Sediment drifts and sediment waves are formed east of the moat. Similar features are observed at the inner, western flank of William’s Ridge thus outlining the recirculation of the AABW branch in the Labuan Basin. The chronological and spatial will be reconstructed via the analysis of those sedimentary structures to provide constraints on climate and ocean circulation variability.

The catalytic performance of nanoparticles (NPs) of Ag anchored on different supports was evaluated during the selective hydrogenation of 1-pentyne and the purification of a mixture of 1-pentene/1-pentyne (70/30 vol %). The catalysts were identified: Ag/Al (Ag supported on ɣ-Al2O3), Ag/Al-Mg (Ag supported on ɣ-Al2O3 modified with Mg), Ag/Ca (Ag supported on CaCO3) and Ag/RX3 (Ag supported on activated carbon-type: RX3). Besides, in situ DRIFTS analysis of 1-pentyne adsorption on each supports, catalysts and 1-pentyne hydrogenation were investigated. The results showed that the synthesized catalysts were active and very selective (≥85%) for obtaining the desired product (1-pentene). Different adsorbed species (-CºC- and -C=C-) were observed on the supports and catalysts surface using in situ DRIFT analysis, which can be correlated to the activity and high selectivity reached. The role of the supports and electronic properties over Ag improve to H2 dissociative chemisorption during the hydrogenation reactions; promoting the selectivity and the high catalytic performance. Ag/Al and Ag/Al-Mg were the most active catalysts. This was due to the synergism between the active Ag°/Ag+ species and the supports (electronic effects). The results show that Ag/Al and Ag/Al-Mg catalysts have favorable properties and are promising for the alkyne hydrogenation and olefin purification reactions.

In this article, we provide various analytical and numerical methods for calculating the average drift of magnetically trapped particles across field lines in complex geometries, and we compare these methods against each other. To evaluate bounce integrals, we use a generalization of the trapezoidal rule which is able to circumvent integrable singularities. We contrast this method with more standard quadrature methods in a parabolic magnetic well and find that the computational cost is significantly lower for the trapezoidal method, though at the cost of accuracy. With numerical routines in place, we next investigate conditions on particles which cross the computational boundary, and we find that important differences arise for particles affected by this boundary, which can depend on the specific implementation of the calculation. Finally, we investigate the bounce-averaged drifts in the optimized stellarator NCSX. From investigating the drifts, one can readily deduce important properties, such as what subset of particles can drive trapped-particle modes and in what regions radial drifts are most deleterious to the stability of such modes.

The prediction of power fluxes and plasma-wall interactions impacted by MHD processes during ITER operation [disruption, Edge Localized Modes (ELMs), 3D magnetic fields applied for ELM control, etc.] requires models that include an accurate description of the MHD processes themselves, as well as of the edge plasma and plasma-wall interaction processes. In this paper, we report progress on improving the edge plasma physics models in the nonlinear extended MHD code JOREK, which has capabilities to simulate the MHD response of the plasma to the applied external 3D fields, disruptions and ELMs. The extended MHD model includes E × B drifts, diamagnetic drifts, and neoclassical flows. These drifts can have large influences, on e.g., divertor asymmetries. Realistic divertor conditions are important for impurity sputtering, transport, and their effect on the plasma. In this work, we implemented kinetic and fluid neutral physics modules, investigated the influence of poloidal flows under divertor conditions in the ITER PFPO-1 (1.8T/5MA) H-mode plasma scenario, and compared the divertor plasma conditions and heat flux to the wall for both the fluid and kinetic neutral model (in JOREK) to the well-established 2D boundary plasma simulation code suite SOLPS-ITER. As an application of the newly developed model, we investigated time-dependent divertor solutions and the transition from attached to partially detached plasmas. We present the formation of a high-field-side high-density-region and how it is driven by poloidal E × B drifts.

Millions of vulnerable consumer IoT devices in home networks are the enabler for cyber crimes putting user privacy and Internet security at risk. Internet service providers (ISPs) are best poised to play key roles in mitigating risks by automatically inferring active IoT devices per household and notifying users of vulnerable ones. Developing a scalable inference method that can perform robustly across thousands of home networks is a non-trivial task. This paper focuses on the challenges of developing and applying data-driven inference models when labeled data of device behaviors is limited and the distribution of data changes (concept drift) across time and space domains. Our contributions are three-fold: (1) We collect and analyze network traffic of 24 types of consumer IoT devices from 12 real homes over six weeks to highlight the challenge of temporal and spatial concept drifts in network behavior of IoT devices; (2) We analyze the performance of two inference strategies, namely "global inference" (a model trained on a combined set of all labeled data from training homes) and "contextualized inference" (several models each trained on the labeled data from a training home) in the presence of concept drifts; and (3) To manage concept drifts, we develop a method that dynamically applies the ``closest'' model (from a set) to network traffic of unseen homes during the testing phase, yielding better performance in 20% of scenarios., Comment: Submitted to IEEE IoT Journal

Satellite data records of stratospheric water vapour have been compared to balloon-borne frost point hygrometer (FP) profiles that are coincident in space and time. The satellite data records of 15 different instruments cover water vapour data available from January 2000 through December 2016. The hygrometer data are from 27 stations all over the world in the same period. For the comparison, real or constructed averaging kernels have been applied to the hygrometer profiles to adjust them to the measurement characteristics of the satellite instruments. For bias evaluation, we have compared satellite profiles averaged over the available temporal coverage to the means of coincident FP profiles for individual stations. For drift determinations, we analysed time series of relative differences between spatiotemporally coincident satellite and hygrometer profiles at individual stations. In a synopsis we have also calculated the mean biases and drifts (and their respective uncertainties) for each satellite record over all applicable hygrometer stations in three altitude ranges (10-30 hPa, 30-100 hPa, and 100 hPa to tropopause). Most of the satellite data have biases <10 % and average drifts <1 % yr-1 in at least one of the respective altitude ranges. Virtually all biases are significant in the sense that their uncertainty range in terms of twice the standard error of the mean does not include zero. Statistically significant drifts (95 % confidence) are detected for 35 % of the ≈ 1200 time series of relative differences between satellites and hygrometers. © Author(s) 2023.

Methanol carbonylation was evaluated over heterogeneous catalysts based on Cu-exchanged zeolitic materials with different topology: Cu@MOR, Cu@FER, and Cu@ZSM-5. Despite the similar Si/Al ratios, it is crucial to acknowledge that the acid strength is influenced by the framework topology, as supported by the NH3-TPD results. This, along with other characterization techniques allowed us to estimate the impact of pore size and pore distribution in these microporous materials on catalytic performance. The channel structure influenced catalytic parameters such as conversion and selectivity. The higher methanol conversion achieved on Cu@FER shows the importance of Brønsted acid sites and redox centres location regarding the topology of the material. Concerning the selectivity, the production of acetic acid was endorsed by the 12-MR (MOR) channels, methyl acetate's production by the 10-MR (FER) channels. Finally, the presence of 6-MR (ZSM-5) channels led to a complete selectivity towards DME production. The reaction mechanism was elucidated via operando DRIFTS-MS and results revealed a bifunctional mechanism in which methanol adsorbs and dehydrates on acidic Brønsted sites and CO is activated over Cu+ species.

Methanol carbonylation was evaluated over heterogeneous catalysts based on Cu-exchanged zeolitic materials with different topology: Cu@MOR, Cu@FER, and Cu@ZSM-5. Despite the similar Si/Al ratios, it is crucial to acknowledge that the acid strength is influenced by the framework topology, as supported by the NH3-TPD results. This, along with other characterization techniques allowed us to estimate the impact of pore size and pore distribution in these microporous materials on catalytic performance. The channel structure influenced catalytic parameters such as conversion and selectivity. The higher methanol conversion achieved on Cu@FER shows the importance of Brønsted acid sites and redox centres location regarding the topology of the material. Concerning the selectivity, the production of acetic acid was endorsed by the 12-MR (MOR) channels, methyl acetate's production by the 10-MR (FER) channels. Finally, the presence of 6-MR (ZSM-5) channels led to a complete selectivity towards DME production. The reaction mechanism was elucidated via operando DRIFTS-MS and results revealed a bifunctional mechanism in which methanol adsorbs and dehydrates on acidic Brønsted sites and CO is activated over Cu+ species.

Continuous machine learning pipelines are common in industrial settings where models are periodically trained on data streams. Unfortunately, concept drifts may occur in data streams where the joint distribution of the data X and label y, P(X, y), changes over time and possibly degrade model accuracy. Existing concept drift adaptation approaches mostly focus on updating the model to the new data possibly using ensemble techniques of previous models and tend to discard the drifted historical data. However, we contend that explicitly utilizing the drifted data together leads to much better model accuracy and propose Quilt, a data-centric framework for identifying and selecting data segments that maximize model accuracy. To address the potential downside of efficiency, Quilt extends existing data subset selection techniques, which can be used to reduce the training data without compromising model accuracy. These techniques cannot be used as is because they only assume virtual drifts where the posterior probabilities P(y|X) are assumed not to change. In contrast, a key challenge in our setup is to also discard undesirable data segments with concept drifts. Quilt thus discards drifted data segments and selects data segment subsets holistically for accurate and efficient model training. The two operations use gradient-based scores, which have little computation overhead. In our experiments, we show that Quilt outperforms state-of-the-art drift adaptation and data selection baselines on synthetic and real datasets., Comment: Accepted to AAAI 2024