Core Model Research Articles

The Preparation of an Ultrafine Copper Powder by the Hydrogen Reduction of an Ultrafine Copper Oxide Powder and Reduction Kinetics.

Ultrafine copper powders were prepared by the air-jet milling of copper oxide (CuO) powders and a subsequent hydrogen (H2) reduction. After milling, the particle size and grain size of CuO powders decreased, while the specific surface area and structural microstrain increased, thereby improving the reaction activity. In a pure H2 atmosphere, the process of CuO reduction was conducted in one step, and followed a pseudo-first-order kinetics model. The smaller CuO powders after milling exhibited higher reduction rates and lower activation energies compared with those without milling. Based on the unreacted shrinking core model, the reduction of CuO powders via H2 was controlled by the interface reaction at the early stage, whereas the latter was limited by the diffusion of H2 through the solid product layer. Additionally, the scanning electron microscopy (SEM) indicated that copper powders after H2 reduction presented a spherical-like shape, and the sintering and agglomeration between particles occurred after 300 °C, which led to a moderate increase in particle size. The preparing parameters (at 400 °C for 180 min) were preferred to obtain ultrafine copper powders with an average particle size in the range of 5.43-6.72 μm and an oxygen content of less than 0.2 wt.%.

Packing Soft Polygons in a Minimum Height Rectangular Target Domain

The paper studies packing polygons of variable shapes, regarding the stretching coefficient, in a rectangular target domain of minimum height. Packing objects of a variable shape have a wide spectrum of applications, e.g, in biology, materials science, mechanics, land allocation, and logistics. Interest in these problems is also due to the modeling of the structures of porous media under pressure, e.g., for creating test models of artificial digital cores. Elements of porous media can be deformed under the influence of an external force, but the mass of each particle remains unchanged. This corresponds to conservation of area for the two-dimensional case. Polygonal objects must be completely contained within the target domain (containment constraint) and do not overlap (non-overlapping constraint), provided they have free translations, continuous rotations, stretch transformations, and conserve their area. The phi-function technique is used for an analytical description of the placement constraints for variable shape polygons. Quasi-phi-functions for describing non-overlapping constraints and phi-functions for describing containment constraints are defined. The packing problem is presented in the form of a nonlinear programming model. A solution strategy is proposed, which consists of the following stages: generation of feasible starting points; search for local minima of the problem of packing soft polygons for each starting point using the decomposition algorithm; choosing the best local minimum found at the previous stage. To search for smart starting arrangements, an optimization algorithm for packing original polygons using their homothetic transformations is applied. Decomposition of the problem of packing polygons of variable shapes is based on an iterative procedure that allows reducing a large-scale problem to a sequence of smaller nonlinear programming problems (linear to the number of objects). Numerical examples are provided for oriented rectangles and non-oriented regular polygons. Keywords: packing, polygons, stretch transformation, rectangular container, quasi-phi-functions, optimization, decomposition algorithm.

High thermal conductivity of the Earth's core and geodynamo

The geomagnetic field is generated by dynamo processes in the Earth's core. This process is supported by the cooling of the planet and the growth of the solid core. It is known that the magnetic field existed long before the appearance of the solid core, although according to available estimates and calculations, this could not have happened. Moreover, according to some models, thermal conductivity in the Earth's core may be three times higher than generally accepted values. In this case, the magnitude of the convective heat flow decreases and the generation of the magnetic field stops. In the above-mentioned models it was assumed that the magnitude of the heat flow at the core – mantle boundary over the entire existence of the Earth linearly decreased by only 15–20 %. The latter is a rough guess. Significantly large changes in heat flow are predicted by models of mantle cooling, in which the growth of convective heat flow in ancient times occurs due to a decrease in the viscosity of the mantle with an increase in the temperature of the substance. Below we consider a model of the cooling of the Earth with a threefold value of the thermal conductivity of the core. It is shown that the use of a combined model of core and mantle cooling makes it possible to significantly increase the heat flow at the core – mantle boundary in ancient times. To slow down the growth of the solid core, a subadiabatic layer is included in the model. As a result, the size of the solid core in the model satisfies seismological observations. The model allows us to obtain a sufficient amount of energy to generate a geomagnetic field, starting from the moment the liquid core appeared and to the present. The appearance of a solid core 2.4 billion years ago, which does not lead to sharp changes in heat flow in the model, is consistent with paleomagnetic observations that do not record changes in the behavior of the magnetic field. The model does not exclude the existence of a multipole magnetic field at the initial stage.

Dissolution behavior and kinetics of copper slag under oxidative conditions

This study investigates the extraction and leaching kinetics of metals from slag, a by-product in pyrometallurgical copper production, under atmospheric leaching conditions in the presence of HCl-H2O2.The optimum leaching conditions were determined as follows: 0.1 M HCl concentration, 3 M H2O2 concentration, 60 minutes leaching time, 318 K leaching temperature, 400 rpm stirring speed, and 25:1 liquid-solid ratio. Under these conditions, 76% Cu, 2.8% Fe, and 23.6% Zn were successfully extracted. The selection of these optimum conditions aimed to achieve a selective leaching outcome by considering the Cu/Fe ratio.The leaching kinetics were found to be in agreement with a mixed control model based on reactant concentrations, as described by the equation (15(1−∝)−53−141−∝−43+120=kt). The shrinking core model was used to represent the leaching process. The activation energy was determined to be 55.35 kJ/mol, and the reaction degree with respect to H2O2 and HCl was found to be 0.90 and 0.58, respectively.SEM-EDX, XRD, and particle size analysis were conducted on the leach residue. The results revealed the presence of Fe2SiO4 and Mg0.2Fe2.6Al0.2O4 phases in the leach balance, while the particle size was determined to be D(50)= 52.5 µm.

Mechanical quenching phenomenon in diamond

The structure of dislocation cores, the fundamental knowledge on crystal plasticity, remains largely unexplored in covalent crystals. Here, we conducted atomically resolved characterizations of dislocation core structures in a plastically deformed diamond anvil cell tip that was unloaded from an exceptionally high pressure of 360 GPa. Our observations unveiled a series of nonequilibrium dislocation cores that deviate from the commonly accepted "five-seven-membered ring" dislocation core model found in FCC-structured covalent crystals. The nonequilibrium dislocation cores were generated through a process known as "mechanical quenching," analogous to the quenching process where a high-energy state is rapidly frozen. The density functional theory-based molecular dynamic simulations reveal that the phenomenon of mechanical quenching in diamond arises from the challenging relaxation of the nonequilibrium configuration, necessitating a large critical strain of 25% that is difficult to maintain. Further electronic-scale analysis suggested that such large critical strain is spent on the excitation of valance electrons for bond breaking and rebonding during relaxation. These findings establish a foundation for the plasticity theory of covalent materials and provide insights into the design of electrical and luminescent properties in diamond, which are intimately linked to the dislocation core structure.

Concussion history and virtual reality metrics predict core or lower extremity injury occurrence among high school athletes.

A history of concussion is recognized as a risk factor for musculoskeletal injury, which is likely associated with physiological effects that warrant better understanding. This study aimed to assess the potential of measurements obtained from an immersive virtual reality (VR) test to identify a subtle perceptual-motor impairment that may be prospectively associated with the occurrence of a core or lower extremity sprain or strain. A cohort of 68 high school athletes (41 female soccer players and 27 male football players) provided survey responses and completed an immersive VR test several days prior to the initiation of preseason practice sessions. Measurements of eye, neck, arm, and whole-body displacements were obtained during 40 successive lunging/reaching responses to visual stimuli moving horizontally across the VR headset display. Injury occurrences were electronically documented from the initial preseason practice session to the final game of the season. A statistically significant and intrinsically credible two-factor prediction model for core or lower extremity injury occurrence included an interaction between female sex and a self-reported history of two or more concussions, along with slow response time (RT) for arm reach (OR = 4.67; 95% CI, 1.51-14.43). Follow-up analyses identified sex-specific cut points for arm reach RT associated with elevated injury risk, which were ≥1.385 s for females and ≥1.257 s for males. High school female soccer players who have sustained more than one concussion appear to be highly vulnerable to core or lower extremity sprain or strain, with the risk of injury compounded by a slow arm reach RT. Male football players as a group demonstrated significantly faster arm reach RT than that of female soccer players, but slow perceptual-motor RT for arm reach was also identified as a potentially important injury risk factor for male players. Immersive VR appears to provide precise measurements of behavioral performance characteristics that depend on brain processing efficiency. Given that the speed, accuracy, and consistency of perceptual-motor responses may be modifiable, future research should explore the potential benefits of VR training for reducing the risk of sport-related injuries.

Multidimensional Impact of Dupilumab on Chronic Rhinosinusitis with Nasal Polyps: A Complete Health Technology Assessment of Clinical, Economic, and Non-Clinical Domains.

Chronic RhinoSinusitis with Nasal Polyps (CRSwNP) represents a condition mainly caused by the type 2 inflammation presence and marked by the existence of polyps within the nasal and paranasal sinuses. The standard of care includes intranasal steroids, additional burst of systemic steroids, if needed, and surgery. However, recurrence is common, especially among patients with comorbid type 2 inflammatory diseases. Recently, biological drugs, addressing the underlying cause of the disease, have been approved in Italy (dupilumab, omalizumab, and mepolizumab). A Health Technology Assessment was conducted to define multidimensional impact, assuming Italian NHS perspective and a 12-month time horizon. The EUnetHTA Core Model was deployed, using the following methods to analyze the domains: (i) literature evidence; (ii) administration of semi-structured questionnaires to 17 healthcare professionals; (iii) health economics tools to define the economic sustainability for the system. Evidence from NMA and ITC showed a more favorable safety profile and better efficacy for dupilumab compared with alternative biologics. All the analyses, synthesizing cost and efficacy measures, showed that dupilumab is the preferable alternative. Specifically, the cost per responder analysis for dupilumab, exhibiting a 67.0% response rate at Week 52, is notably economical at 14,209EUR per responder. This presents a more economical profile compared with the cost per responder for omalizumab (36.2% response rate) at 24,999EUR and mepolizumab (28.5% response rate) at 31,863EUR. These results underscore dupilumab's potential, not merely in terms of clinical outcomes, but also in terms of economic rationality, thereby solidifying its status as a valid and preferrable alternative in the management of CRSwNP, in the context of the Italian NHS.

Equilibrium core modeling of a pebble bed reactor similar to the Xe-100 with SCALE

As the nuclear industry moves towards licensing and constructing advanced reactors, new attention has been focused on the advanced reactor designs that have past operational experience, such as pebble-bed high-temperature gas-cooled reactors (PB-HTGRs). Pebble-bed reactor designs have many advantages, such as their higher operating temperatures and online refueling capabilities. However, high-fidelity computational modeling of pebble-bed reactor designs, from reactor startup to operation at equilibrium, is more challenging compared to conventionally fueled reactors due to the continuous movement of the fuel pebbles through the reactor during operation. In previous work at Oak Ridge National Laboratory (ORNL), the SCALE Leap-In method for Cores at Equilibrium (SLICE) was developed around tools within the SCALE code system. This iterative method can effectively generate pebble-bed reactor zone-wise fuel inventories at equilibrium core operation within a reasonable computational time. The objective of this work was to further verify the ORNL SLICE method and to investigate the impact of considering temperature profiles during the application of the method. The SLICE method was applied to a modular high-temperature gas-cooled reactor design based upon publicly available design specifications of the Xe-100 pebble-bed reactor. Upon comparing the results from the SLICE method to published literature, the differences in the eigenvalue keffective were on the order of several hundred pcm (percent millirho). To investigate one possible cause of these differences, a study looking at the sensitivity of the full-core equilibrium keffective and discharge nuclide inventory to temperature was performed by developing equilibrium cores of two additional temperature profiles. From this temperature study, differences on the order of hundreds of pcm for the full-core equilibrium keffective, and up to 15% difference for the discharge inventories were found. These results indicated the strong dependence on temperature that needs to be considered for future work in equilibrium modeling of PB-HTGRs.

Buckling and post-buckling of anisogrid lattice-core sandwich plates with nanocomposite skins

This research offers an in-depth examination of the buckling and post-buckling responses observed in sandwich plates distinguished by the presence of an anisogrid lattice core and face layers composed of functionally graded carbon nanotube-reinforced composites (FG-CNTRC). To model this structure, a global continuous model for the lattice core is employed, while the displacement field is approximated using the First-Order Shear Deformation Theory (FSDT). Geometrically nonlinear strain relations based on the von-Kármán assumptions are incorporated to accurately capture the structural response. The homogenization of the FG-CNTRC skins is achieved through the modified rule of mixture, enabling estimations of effective material properties. To separate pre-buckling and post-buckling phases, we implement the adjacent-equilibrium criterion. Solving the linear (pertaining to buckling) and nonlinear (related to post-buckling) equations is based on the Generalized Differential Quadrature (GDQ) technique alongside an iterative method grounded in the displacement control strategy. Our analysis explores the influence of core and skin characteristics on the stability of these sandwich plates. By examining critical buckling thresholds, post-buckling equilibrium paths, and the underlying mechanisms driving these responses, we contribute insights into the structural stability and performance of such composite structures. This study not only enhances our understanding of buckling behavior in anisogrid lattice core sandwich plates but also holds promise for diverse engineering applications where structural stability is paramount.

Design of Janus Nanoparticles for Mobility Control in Heterogeneous Reservoirs.

Unfavorable mobility ratios in heterogeneous reservoirs have resulted in progressively poor waterflood sweep efficiency and diminishing production. In order to address this issue, our study has developed amphiphilic-structured nanoparticles aimed at enhancing the microscopic displacement capability and oil displacement efficiency. First, the transport process of Janus nanoparticles in porous media was investigated. During the water flooding, Janus nanoparticle injection, and subsequent water flooding stages, the injection pressure increased in a "stepped" pattern, reaching 0.023, 0.029, and 0.038 MPa, respectively. Second, emulsification effects and emulsion viscosity experiments demonstrated that the amphiphilic structure improved the interaction at the oil-water interface, reducing the seepage resistance of the oil phase through emulsification. In porous media, Janus nanoparticles transported with water exhibit 'self-seeking oil' behavior and interact with the oil phase, reducing the viscosity of the oil phase from 19 to 5 mPa·s at 80 °C. Finally, the core model displacement experiment verified the characteristics of Janus nanoparticles in improving the oil-water mobility ratio. Compared with the water flooding stage, the recovery percent increased by 20.8%, of which 13.7% was attributed to the subsequent water flooding stage. Utilizing the asymmetry of the Janus particle structure can provide an effective path to enhanced oil recovery in inhomogeneous reservoirs.

MgAl2O4–C refractories with balanced properties for a delivery device used in the double-roll thin strip continuous casting

To realize the double-roll thin strip continuous casting process for the production of high-quality steels, it is necessary to seek a new type of refractory material tailored for delivery device employed in this process. Therefore, the present study aimed to design and prepare MgAl2O4–C refractories for this purpose as well as to study their physical and mechanical properties, thermal shock resistance, oxidation kinetics, and to perform molten steel experiments after pre-oxidation treatment. The results of the present study demonstrated that the specimens could maintain high mechanical and thermal shock properties even after pre-oxidation treatment. After pre-oxidation at 1100 °C/1 h, the cold compressive strength of the specimen reached 17.63 MPa, with a residual strength ratio (thermal shock resistance) of 77.99%. Analysis of oxidation kinetics based on the shrinking core model indicated an increase in the effective diffusion coefficient and complete oxidation time with higher pre-oxidation temperature. Pre-oxidation of MgAl2O4–C refractories has been shown to be effective in reducing carbon penetration into the molten steel from carbon-containing refractories, thus maintaining the stability of the molten steel composition. This effect was attributed to enhanced contact and wettability between the steel and refractories, an augmented thickness of the pre-oxidized decarburized layer, the formation of a dense magnesium oxide layer, and the emergence of the MgAl1·9Fe1O4 phase. Consequently, an MgAl2O4–C refractories designed for the delivery device was proposed along with anticipated applicability in the double-roll continuous casting process.

An Application of the Core and Balance Model of Family Leisure Functioning to Study Leisure Activities Practised by Families from Wielkopolska

The purpose of this study was to examine family leisure involvement in 60 Polish families from Wielkopolska from the perspective of parents and their children. The Core and Balance Model of Family Leisure Functioning (CBM) was used to examine the relationship between core and balance activities and the level of satisfaction with family leisure involvement. Two questionnaires were used: the Family Leisure Activity Profile (FLAP) and the Family Leisure Satisfaction Scale (FLSS). The families in the sample were found to take part in core activities more frequently. While the level of involvement in family leisure was average, the level of satisfaction derived from it was relatively high. The findings indicate that the Polish version of the FLAP can be a useful tool in the study of leisure behaviour of Polish families.

Kinetics of Estonian Phosphate Rock Dissolution in Hydrochloric Acid

The kinetics of the dissolution of Estonian phosphate rock and the governing reaction mechanisms in hydrochloric acid in technological processes were investigated. The influences of particle size and acid concentrations of 0.5–1.5 M on the reaction rate and the pH variation during the process were studied at a dosage of 2.1 moles of HCl per mole of calcium for 60 min. The results indicated that the solubility of phosphorus reached 94%–100% for the fine samples and 82%–99% for the coarse samples. The time required to achieve an apparent steady-state pH reduced with the increasing acid concentrations and decreasing particle sizes. It was determined that the CaF2 precipitation in solutions starting at 1 M was faster at higher concentrations. The SEM surface analysis of the insoluble particles proved the existence of etch pit formation. The XPS and EDX analyses affirmed that the dissolution was incongruent. The surface composition of the unreacted particles gave a stoichiometry of CaF1.8, showing the formation of CaF2 on the surface. The dissolution kinetics were analyzed using the shrinking core model and showed a combination of chemical reaction, diffusion or interfacial transfer, and diffusion, sequentially for coarse particles or simultaneously for fine fractions.

Evaluation of Permeability Models for Foundry Molds and Cores in Sand Casting Processes

Predicting the permeability of different regions of foundry cores and molds with complex geometries will help control the regional outgassing, enabling better defect prediction in castings. In this work, foundry cores prepared with different bulk properties were characterized using X-ray microtomography, and the obtained images were analyzed to study all relevant grain and pore parameters, including but not limited to the specific surface area, specific internal volume, and tortuosity. The obtained microstructural parameters were incorporated into prevalent models used to predict the fluid flow through porous media, and their accuracy is compared with respect to experimentally measured permeability. The original Kozeny model was identified as the most suitable model to predict the permeability of sand molds. Although the model predicts permeability well, the input parameters are laborious to measure. Hence, a methodology for replacing the pore diameter and tortuosity with simple process parameters is proposed. This modified version of the original Kozeny model helps predict permeability of foundry molds and cores at different regions resulting in better defect prediction and eventual scrap reduction.

Analysis of single particle and collective beam effects in high intensity beams in a periodic quadrupole channel

In high intensity proton accelerators, there are two main mechanisms that can cause beam degradation: incoherent and coherent effects due to nonlinear space charge forces. The incoherent effects represent the single particle dynamics while the coherent effects represent the collective response of the beam. Of particular interest is the region above a zero current phase advance (σ0) of 90° where the (coherent) second-order envelope instability and the (incoherent) fourth-order particle resonance are seen to lead to emittance growth. Large emittance growth is also seen below the envelope instability region as the full current beam phase advance (σ) decreases. In the present study, we have studied the nonlinear effects in a high intensity beam propagating through a focusing-defocusing (FD) quadrupole channel for σ0 greater than 90°, both analytically, by studying the solutions of Kapchinsky-Vladimirsky (KV) equation and the particle core model, and through detailed particle-in-cell (PIC) simulations using the code. The KV envelope equation gives the collective response of the beam while the particle core model gives the contribution of the single particle effects. With pic simulations, which resemble the behavior of the real beams more closely as compared to envelope calculations or the particle core model, it is possible to study the evolution of the beam in a self-consistent manner. Our studies show, that it is possible to identify the specific process responsible for beam degradation in high intensity beams. It is also possible to identify which process dominates under different conditions. We further show that the width of the emittance increase stop band calculated from PIC simulations is wider than that calculated by the envelope equations and that the width depends on the length of the channel being studied. Published by the American Physical Society 2024

Aligning open educational resources to new taxonomies: How AI technologies can help and in which scenarios

Aligning open educational resources (OER) to skill taxonomies is a common task in the education field and helps teachers better locate material that aligns with the standards of their curriculum. When taxonomies change, as they periodically do, re-tagging the increasing mass of open educational resources is needed. The process of manual tagging is, however, exceedingly labor intensive. We propose and evaluate a novel combination of machine learning methods to help automate tagging open educational resources with skills from an existing taxonomy as well as skills from any newly introduced taxonomy. We collected text, image figures, and videos from tens of thousands of educational resources from two major digital learning platforms to answer the research questions of: how effective are machine learning models in automatically updating OER classification to reflect a new taxonomy (RQ1), and which models may be of practical use in different scenarios (RQ2)? Using several taxonomies, including the US Common Core, we find that while full automation is not practically viable, our most generalizable model can reach non-expert human labeling performance requiring only 100 labeled examples and near expert level with 5000. We believe these novel findings may have immediate utility for practitioners and policymakers and better ready the growing landscape of open educational resources for the advent of new taxonomies ahead. We publicly release our pre-trained US Common Core and new taxonomy tagging models, providing guidance on their viability in various real-world scenarios.

Absorption Distribution Metabolism Excretion and Toxicity Property Prediction Utilizing a Pre-Trained Natural Language Processing Model and Its Applications in Early-Stage Drug Development.

Machine learning techniques are extensively employed in drug discovery, with a significant focus on developing QSAR models that interpret the structural information of potential drugs. In this study, the pre-trained natural language processing (NLP) model, ChemBERTa, was utilized in the drug discovery process. We proposed and evaluated four core model architectures as follows: deep neural network (DNN), encoder, concatenation (concat), and pipe. The DNN model processes physicochemical properties as input, while the encoder model leverages the simplified molecular input line entry system (SMILES) along with NLP techniques. The latter two models, concat and pipe, incorporate both SMILES and physicochemical properties, operating in parallel and with sequential manners, respectively. We collected 5238 entries from DrugBank, including their physicochemical properties and absorption, distribution, metabolism, excretion, and toxicity (ADMET) features. The models' performance was assessed by the area under the receiver operating characteristic curve (AUROC), with the DNN, encoder, concat, and pipe models achieved 62.4%, 76.0%, 74.9%, and 68.2%, respectively. In a separate test with 84 experimental microsomal stability datasets, the AUROC scores for external data were 78% for DNN, 44% for the encoder, and 50% for concat, indicating that the DNN model had superior predictive capabilities for new data. This suggests that models based on structural information may require further optimization or alternative tokenization strategies. The application of natural language processing techniques to pharmaceutical challenges has demonstrated promising results, highlighting the need for more extensive data to enhance model generalization.

Leaching of Nickel and Cobalt from a Mixed Nickel-Cobalt Hydroxide Precipitate Using Organic Acids

Nickel (Ni) and cobalt (Co) are strategic metals that have found applications in a wide range of metallurgical and industrial uses. In this study, the dissolution of a mixed nickel–cobalt hydroxide precipitate using organic acids (citric, oxalic, and malic acid) was investigated. Citric acid was found to be the best leaching agent yielding the following dissolution rates: 91.2% Ni, 86.8% Co, and 90.8% Mn. Oxalic acid resulted in low dissolution, which is likely due to the formation of insoluble metal oxalates. The impact of acid concentration, leaching time, and temperature on metal dissolution was systematically examined. The optimal dissolution conditions were identified as 0.5 M citric acid at 30 °C for 30 min, utilizing a 1/20 solid/liquid ratio and a stirring speed of 400 revolutions per minute (rpm). The attempt to use oxidants, such as potassium permanganate (KMnO4) and hydrogen peroxide (H2O2), to achieve selective dissolution in an organic acid environment was not successful, which was different from that in the sulfuric acid case. As for the leaching kinetics in the organic acids, it seems that the leaching of Ni correlates with the Shrinking Core Model, specifically regarding porous-layer diffusion control. Based on the experimental results, the activation energy for the leaching of Ni was estimated to be 3.1 kJ/mol.

Study on the pyrolysis characteristics and char gasification kinetics of pre-separated automobile shredder residues

Automobile shredder residue (ASR) is a typical solid waste produced by the end-of-life automobile industry. The pyrolysis and gasification of ASR are effective methods to achieve reduction and resource utilization. By means of thermogravimetric analysis and fixed-bed experiments, the distribution as well as physicochemical properties of five typical ASR pyrolysis products after pre-sorting were systematically investigated, and the gasification kinetics of their pyrolysis char under CO2 as well as steam were also studied. Results showed that the pyrolysis of ASR was basically completed at 500°C, and the pyrolysis gas yield increased significantly with temperature. The tar contained mainly poly-benzene rings with carbon atoms C11–15 and their derivatives. At high temperatures, the structure of pyrolysis char tended to become more ordered, with small functional group such as - OH, - NH, - CH3, - CC -, and - COOH fracturing. The residual amounts of heavy metals such as Cr, Pb, and Ni in pyrolysis char were mostly exceed 45%, which were significantly affected by the type of heavy metals, pyrolysis temperature and material type. Hydrogen production from steam gasification accounted for up to 65%, and the gasification reactivity was better than that of CO2 gasification. The activation energy (E) calculated through shrinking core model (SCM) and random pore model (RPM) was 38.81–105.44 kJ/mol and 39.53–98.56 kJ/mol, respectively, which provided a good fit to the gasification kinetics in two gasification atmospheres. This study provides detailed data for the ASR pyrolysis and gasification technology, which is an important reference value for ASR research.

Temporal teacher with masked transformers for semi-supervised action proposal generation

By conditioning on unit-level predictions, anchor-free models for action proposal generation have displayed impressive capabilities, such as having a lightweight architecture. However, task performance depends significantly on the quality of data used in training, and most effective models have relied on human-annotated data. Semi-supervised learning, i.e., jointly training deep neural networks with a labeled dataset as well as an unlabeled dataset, has made significant progress recently. Existing works have either primarily focused on classification tasks, which may require less annotation effort, or considered anchor-based detection models. Inspired by recent advances in semi-supervised methods on anchor-free object detectors, we propose a teacher-student framework for a two-stage action detection pipeline, named Temporal Teacher with Masked Transformers (TTMT), to generate high-quality action proposals based on an anchor-free transformer model. Leveraging consistency learning as one self-training technique, the model jointly trains an anchor-free student model and a gradually progressing teacher counterpart in a mutually beneficial manner. As the core model, we design a Transformer-based anchor-free model to improve effectiveness for temporal evaluation. We integrate bi-directional masks and devise encoder-only Masked Transformers for sequences. Jointly training on boundary locations and various local snippet-based features, our model predicts via the proposed scoring function for generating proposal candidates. Experiments on the THUMOS14 and ActivityNet-1.3 benchmarks demonstrate the effectiveness of our model for temporal proposal generation task.