Clothing Parsing Based on Multi-Scale Fusion and Improved Self-Attention Mechanism

Due to the lack of long-range association and spatial location information,fine details and accurate boundaries of complex clothing images cannot always be obtained by using the existing deep learning-based methods.This paper presents a convolutional structure with multi-scale fusion to optimize the step of clothing feature extraction and a self-attention module to capture long-range association information.The structure enables the self-attention mechanism to directly participate in the process of information exchange through the down-scaling projection operation of the multi-scale framework.In addition,the improved self-attention module introduces the extraction of 2-dimensional relative position information to make up for its lack of ability to extract spatial position features from clothing images.The experimental results based on the colorful fashion parsing dataset(CFPD)show that the proposed network structure achieves 53.68%mean intersection over union(mIoU)and has better performance on the clothing parsing task.

Lightweight Cross-Modal Multispectral Pedestrian Detection Based on Spatial Reweighted Attention Mechanism

Multispectral pedestrian detection technology leverages infrared images to provide reliable information for visible light images, demonstrating significant advantages in low-light conditions and background occlusion scenarios. However, while continuously improving cross-modal feature extraction and fusion, ensuring the model’s detection speed is also a challenging issue. We have devised a deep learning network model for cross-modal pedestrian detection based on Resnet50, aiming to focus on more reliable features and enhance the model’s detection efficiency. This model employs a spatial attention mechanism to reweight the input visible light and infrared image data, enhancing the model’s focus on different spatial positions and sharing the weighted feature data across different modalities, thereby reducing the interference of multi-modal features. Subsequently, lightweight modules with depthwise separable convolution are incorporated to reduce the model’s parameter count and computational load through channel-wise and point-wise convolutions. The network model algorithm proposed in this paper was experimentally validated on the publicly available KAIST dataset and compared with other existing methods. The experimental results demonstrate that our approach achieves favorable performance in various complex environments, affirming the effectiveness of the multispectral pedestrian detection technology proposed in this paper.

Remaining Useful Life Prediction of Rail Based on Improved Pulse Separable Convolution Enhanced Transformer Encoder

In order to prevent possible casualties and economic loss, it is critical to accurate prediction of the Remaining Useful Life (RUL) in rail prognostics health management. However, the traditional neural networks is difficult to capture the long-term dependency relationship of the time series in the modeling of the long time series of rail damage, due to the coupling relationship of multi-channel data from multiple sensors. Here, in this paper, a novel RUL prediction model with an enhanced pulse separable convolution is used to solve this issue. Firstly, a coding module based on the improved pulse separable convolutional network is established to effectively model the relationship between the data. To enhance the network, an alternate gradient back propagation method is implemented. And an efficient channel attention (ECA) mechanism is developed for better emphasizing the useful pulse characteristics. Secondly, an optimized Transformer encoder was designed to serve as the backbone of the model. It has the ability to efficiently understand relationship between the data itself and each other at each time step of long time series with a full life cycle. More importantly, the Transformer encoder is improved by integrating pulse maximum pooling to retain more pulse timing characteristics. Finally, based on the characteristics of the front layer, the final predicted RUL value was provided and served as the end-to-end solution. The empirical findings validate the efficacy of the suggested approach in forecasting the rail RUL, surpassing various existing data-driven prognostication techniques. Meanwhile, the proposed method also shows good generalization performance on PHM2012 bearing data set.

Keyphrase Generation Based on Self-Attention Mechanism

Keyphrase greatly provides summarized and valuable information.This information can help us not only understand text semantics,but also organize and retrieve text content effectively.The task of automatically generating it has received considerable attention in recent decades.From the previous studies,we can see many workable solutions for obtaining keyphrases.One method is to divide the content to be summarized into multiple blocks of text,then we rank and select the most important content.The disadvantage of this method is that it cannot identify keyphrase that does not include in the text,let alone get the real semantic meaning hidden in the text.Another approach uses recurrent neural networks to generate keyphrases from the semantic aspects of the text,but the inherently sequential nature precludes parallelization within training examples,and distances have limitations on context dependencies.Previous works have demonstrated the benefits of the self-attention mechanism,which can learn global text dependency features and can be parallelized.Inspired by the above observation,we propose a keyphrase generation model,which is based entirely on the self-attention mechanism.It is an encoder-decoder model that can make up the above disadvantage effectively.In addition,we also consider the semantic similarity between keyphrases,and add semantic similarity processing module into the model.This proposed model,which is demonstrated by empirical analysis on five datasets,can achieve competitive performance compared to baseline methods.

Separation of atropisomers by chiral liquid chromatography and thermodynamic analysis of separation mechanism

In the pharmaceutical industry, the analysis of atropisomers is of considerable interest from both scientific and regulatory perspectives. The compound of interest contains two stereogenic axes due to the hindered rotation around the single bonds connecting the aryl groups, which results in four potential configurational isomers(atropisomers). The separation of the four atropisomers was achieved on a derivatized β-cyclodextrin bonded stationary phase. Further investigation showed that low temperature conditions, including sample preparation(-70 °C), sample storage(-70 °C), and chromatographic separation(6 °C), were critical to preventing interconversion. LC-UV-laser polarimetric analysis identified peaks 1 and 2 as a pair of enantiomers and peaks3 and 4 as another. Thermodynamic analysis of the retention data indicated that the separation of the pairs of enantiomers is primarily enthalpy controlled as indicated by the positive slope of the van't Huff plot. The difference in absolute Δ(Δ H), ranged from 2.20 k J/mol to 2.42 k J/mol.

3D inner-outer asymmetric sponge for enormous-volume emulsion wastewater treatment based on a new“demulsification-transport”mechanism

Although oily wastewater treatment realized by superwetting materials has attracted heightened attention in recent years,how to treat enormous-volume emulsion wastewater is still a tough problem,which is ascribed to the emulsion accumulation.Herein,to address this problem,a material is presented by subtly integrating chemical demulsification and 3D inner-outer asymmetric wettability to a sponge substrate,and thus wettability gradient-driven oil directional transport for achieving unprecedented enormous-volume emulsion wastewater treatment is realized based on a“demulsification-transport”mechanism.The maximum treatment volume realized by the sponge is as large as 3 L(2.08×10^(4) L per cubic meter of the sponge)in one cycle,which is about 100 times of the reported materials.Besides,owing to the large pore size of the sponge,9000 L m^(2)h^(-1)(LMH)separation flux and 99.5%separation efficiency are realized simultaneously,which overcomes the trade-off dilemma.Such a 3D inner-outer asymmetric sponge displaying unprecedented advantage in the treatment volume can promote the development of the oily wastewater treatment field,as well as expand the application prospects of superwetting materials,especially in continuous water treatment.

Advances in depressants for flotation separation of Cu–Fe sulfide minerals at low alkalinity:A critical review

The flotation separation of Cu–Fe sulfide minerals at low alkalinity can be achieved using selective depressants.In the flotation system of Cu–Fe sulfide minerals,depressants usually preferentially interact with the pyrite surface to render the mineral surface hydrophilic and hinder the adsorption of the collector.This review summarizes the advances in depressants for the flotation separation of Cu–Fe sulfide minerals at low alkalinity.These advances include use of inorganic depressants (oxidants and sulfur–oxygen compounds),natural polysaccharides (starch,dextrin,konjac glucomannan,and galactomannan),modified polymers (carboxymethyl cellulose,polyacrylamide,lignosulfonate,and tricarboxylate sodium starch),organic acids (polyglutamic acid,sodium humate,tannic acid,pyrogallic acid,salicylic acid,and lactic acid),sodium dimethyl dithiocarbamate,and diethylenetriamine.The potential application of specific inorganic and organic depressants in the flotation separation of Cu–Fe sulfide minerals at low alkalinity is reviewed.The advances in the use of organic depressants with respect to the flotation separation of Cu–Fe sulfide minerals are comprehensively detailed.Additionally,the depression performances and mechanisms of different types of organic depressants on mineral surfaces are summarized.Finally,several perspectives on depressants vis-à-vis flotation separation of Cu–Fe sulfide minerals at low alkalinity are proposed.

Actional Mechanism of Trifluoroacetic Acid for the Separation of Biopolymers by Reversed-phase Liquid Chromatography

The present paper covers the actional mechanism of trifluoroacetic acid for the separation of biopolymers investigated by using the parameters of stoichiometric displacement model for retention(SDM-R) in reversed-phase liquid chromatography. It was found that the trifluoroacetic acid(TFA) may participate in, or stimulate the association among displacing agent molecules in mobile phase, and decrease the affinity of both the associate molecules of the displacing agent and the TFA-protein ion-pairing. The former dominates over the separation selectivity of biopolymers as the concentration of TFA is lower than a given value, and the two contrary functions partly offset to each other and the latter dominates as its concentration is greater than the given value.

Mechanism of pool separation and stratification in laser remelting and cladding

The laser remelting with a two-layer material system (upper material was Al-30 % Ti-20 % Ni alloy,substrate was commercial aluminum alloy) and the laser cladding of a commercial 45 steel with copper Powder (including 25%SiC) were carried out using a 2kW continuous CO2 laser. For the case of laser remelting, a upper Pool in the alloying layer and a lower Pool in the substrate separated by the unmelted Al-Ti-Ni alloy were observed. For laser cladding, a stratified Pool was observed, whose top layer was Cu alloy liquid and bottom was Fe alloy liquid. The mechanism of laser Pool separation and stratification is illustrated by numerical calculation of heat transter process of the two-layer system, combining with material physical properties (especially mixed enthalpy). A classification criterion for laser Pool with the two-layer material system has been presented and four types of the laser Pool are divided into unique Pool, separated Pool, mixed Pool and stratified pool,which provides a theoretical basis for obtaining a excellent surface coating.

Liquid phase separating mechanism and preparation techniques of immiscible alloys

Immiscible alloys have attracted growing interest for their valuable physical and mechanical properties. However, their production is difficult because of metallurgical problems in which there is a serious tendency for gravity separation in the region of the miscibility gap. So far the study on the liquid separation mechanism is still one of the important projects in the spatial materials science and the spatial fluid science. The studied results about the liquid phase separating mechanism of immiscible alloys are presented, at the same time the preparation techniques of homogeneous immiscible alloys are summarized, and the existing problems and the related researching areas in the future are also pointed out.

Mechanical mechanism of overlying strata breaking and development of fractured zone during close-distance coal seam group mining

This study mainly investigates the mechanical mechanism of overlying strata breaking and the development of fractured zones during close-distance coal seam group mining in the Gaojialiang coal mine.First,a mechanical model for the second"activation"of broken overlying strata is established,and the related mechanical"activation"conditions are obtained.A recursive formula for calculating the separation distance of overlying strata is deduced.Second,a height determining method for predicting the height of fractured zones during close-distance coal seam group mining is proposed based on two values,namely,the separation distance and ultimate subsidence value of overlying strata.This method is applied to calculate the fractured zone heights in nos.20107 and 20307 mining faces.The calculated results are almost equal to the field observation results.Third,a modified formula for calculating the height of a waterflowing fractured zone is proposed.A comparison of the calculated and observed results shows that the errors are small.The height determining method and modified formula not only build a theoretical foundation for water conservation mining at the Gaojialiang coal mine,but also provide a reference for estimating the height of water-flowing fractured zones in other coal mines with similar conditions.

Synthesis of mechanically robust porous carbon monoliths for CO2 adsorption and separation

Porous carbon materials with developed porosity,high surface area and good thermal-and chemicalresistance are advantageous for gas adsorption and separation.However,most carbon adsorbents are in powder form which exhibit high pressure drop when deployed in practical separation bed.While monolithic carbons have largely addressed the pulverization problem and preserved kinetics and usually suffer from abrasion during multiple adsorption-desorption cycles.Herein,we proposed the designed synthesis of mechanically robust carbon monoliths with hierarchical pores,solid nitrogen-containing framework.The synthesis started with the polymerization of resorcinol and formaldehyde under weakly acidic conditions generated from cyanuric acid,and then an appropriate amount of hexamethylenetetramine(HMTA)was added as a crosslinker to prompt the formation of three dimensional frameworks.After carbonization process,the as-obtained porous carbon monoliths have a high radial compressive strength of 886 N/cm as well as a BET specific surface area of up to 683 m2/g.At approximately 1 bar,the CO2 equilibrium capacities of the monoliths are in the range of 3.1–4.0 mmol/g at 273 K and of 2.3–3.0 mmol/g at 298 K,exhibiting high selectivity for the capture of CO2 over N2 from a stream which consists of 16.7%(v%)CO2 in N2.Meanwhile,they undergo a facile CO2 release in an argon stream at 298 K,indicating a good regeneration capacity.After cycle testing,sieving and regeneration,the adsorbent has no mass loss,compared to that of its fresh counterpart.

Nano-silica-decorated Poly(m-Phenylene Isophthalamide)Separator with Enhanced Mechanical and Electrolyte Wetting Properties for Lithium-Ion Batteries

Heat-resistant poly(m-phenylene isophthalamide)(PMIA)has attracted considerable attention as a novel separator for application in lithium-ion batteries(LIBs);however,its mechanical strength and electrolyte wettability are not ideal.Herein,a nano-silica-decorated poly(m-phenylene isophthalamide)(PMIA@SiO2)separator was fabricated with SiO2 nanoparticles uniformly attached to the pores and pore walls of the PMIA separator.The as-prepared PMIA@SiO2 separator has good mechanical strength(a 16%improvement compared with pristine PMIA)and wettability toward the electrolyte(the contact angle decreases from 34.0°to 23.1°).The PMIA@SiO2 separator also had a high ionic conductivity(0.75 mS/cm)and low interfacial electric resistance(75Ω).The assembled LiCoO2/PMIA@Si O2-liquid electrolyte/Li cell displays good cycle performance with a capacity retention of 88.1%after 50 cycles.Furthermore,the cycling performance and rate capacity rarely changed after high-temperature treatment.Therefore,the nano-silica-decorated PMIA separator is a potential candidate for application in LIBs with high safety.

Comparison of oil displacement mechanisms and performances between continuous and dispersed phase flooding agents

To compare the oil displacement mechanisms and performances of continuous phase flooding agent(traditional polymer solution) and dispersed phase flooding agent(particle-type polymer SMG dispersion), the particle phase separation of SMG dispersion migrating in pores was simulated by using the microfluidic technology. Theoretically guided by the tree fork concentration distribution of red cells in biological fluid mechanics, the concentration distribution mathematical model of SMG in different pores is established. Furthermore, the micro and macro physical simulation experiments of continuous and dispersed phase flooding agents were carried out. The results show that the continuous flooding agent enters all the swept zones and increases the flow resistance in both larger and small pores. On the contrary, the particle phase separation phenomenon occurs during the injection process of dispersed flooding agent. The SMG particles gather in the larger pore to form bridge blinding, and the carrier fluid displace oil in the small pore. Working in cooperation, the SMG particle and carrier fluid drive the residual oil in the low permeability layers step by step and achieve the goal of enhanced oil recovery. The laboratory experimental results indicate that, the oil increment and water reduction effect of dispersed flooding agent is much better than that of continuous flooding agent, which is consistent with the field test results.

Adsorption of oil from waste water by coal:characteristics and mechanism

The work described here was focused on exploring the potential application of coal to purification of oily waste water.Coal was added to oily waste water as an adsorbent and then removed through a flotation process.This allowed economical and highly efficient separation of oil from the waste water.The absorption time,coal type,coal particle size distribution,pH value and oil concentration were investigated.The results indicate that oil absorption by a coal increases for a period of 1.5 h and then gradually tends toward an equilibrium value.It appears that the absorption capacity of anthracite is more than that of lean coal or lignite,given the same coal particle size distribution.The absorption capacity of a coarse coal fraction is less than that of finer coal,given the same of coal type.The absorption capacity of anthracite decreases slightly as the pH increases from 4 to 9.The adsorption of oil on anthracite follows the Freundlich isothermal adsorption law:given initial oil concentrations of 160.5 or 1023.6 mg/L the absorption capacity was 23.8 or 840.0 mg/g.The absorption mechanism consists of two kinds of absorption,a physical process assisted by a chemical one.

MECHANISM AND APPLICATION OF COAL DRY BENEFICIATION WITH AIR-DENSE MEDIUM FLUIDIZED BED

Ｃｏａｌｄｒｙｂｅｎｅｆｉｃｉａｔｉｏｎｗｉｔｈａｉｒ－ｄｅｎｓｅｍｅｄｉｕｍｆｌｕｉｄｉｚｅｄｂｅｄｈａｓｂｅｅｎｅｓｔａｂｌｉｓｈｅｄａｓａｈｉｇｈｅｆｉ－ｃｉｅｎｃｙｄｒｙｓｅｐａｒａｔｉｏｎｔｅｃｈｎｉｑｕｅｗｈｉｃｈａｐｐｌｉｅｓｆｌｕ－ｉ...

An upgraded polymeric composite with interparticle chemical bonding microstructure toward lithium-ion battery separators with enhanced safety and electrochemical performances

A composite separator of SiC/PVDF-HFP was synthesized for lithium-ion batteries with high thermal and mechanical stabilities.Benefiting from the nanoscale,high hardness,and melting point of SiC,SiC/PVDFHFP with highly uniform microstructure was obtained.This polarization caused by barrier penetration was significantly restrained.Due to the Si-F bond between SiC and PVDF-HFP,the structural stability has been obviously enhanced,which could suppress the growth of lithium(Li) dendrite.Furthermore,some 3D reticulated Si nanowires are found on the surface of Li anode,which also greatly inhibit Li dendrites and result in irregular flakes of Li metal.Especially,the shrinkage of 6% SiC/PVDF-HFP at 150℃ is only 5%,which is notably lower than those of PVDF-HFP and Celgard2500.The commercial LiFePO_(4) cell assembled with 6% SiC/PVDF-HFP possesses a specific capacity of 157.8 mA h g^(-1) and coulomb efficiency of 98% at 80℃.In addition,the tensile strength and modulus of 6% SiC/PVDF-HFP could reach 14.6 and 562 MPa,respectively.And a small deformation(1000 nm) and strong deformation recovery are obtained under a high additional load(2.3 mN).Compared with PVDF-HFP and Celgard2500,the symmetric Li cell assembled with 6% SiC/PVDF-HFP has not polarized after 900 cycles due to its excellent mechanical stabilities.This strategy provides a feasible solution for the composite separator of high-safety batteries with a high temperature and impact resistance.

Rational design and low-cost fabrication of multifunctional separators enabling high sulfur utilization in long-life lithium-sulfur batteries

The lithium-sulfur(Li-S)battery with an ultrahigh theoretical energy density has emerged as a promising rechargeable battery system.However,the practical applications of Li-S batteries are severely plagued by the sluggish reaction kinetics of sulfur species and notorious shuttling of soluble lithium polysulfides(LiPSs)intermediates that result in low sulfur utilization.The introduction of functional layers on separators has been considered as an effective strategy to improve the sulfur utilization in Li-S batteries by achieving effective regulation of LiPSs.Herein,a promising self-assembly strategy is proposed to achieve the low-cost fabrication of hollow and hierarchically porous Fe_(3)O_(4)nanospheres(p-Fe_(3)O_(4)-NSs)assembled by numerous extremely-small primary nanocrystals as building blocks.The rationally-designed p-Fe_(3)O_(4)-NSs are utilized as a multifunctional layer on the separator with highly efficient trapping and conversion features toward LiPSs.Results demonstrate that the nanostructured p-Fe_(3)O_(4)-NSs provide chemical adsorption toward LiPSs and kinetically promote the mutual transformation between LiPSs and Li_(2)S_(2)/Li_(2)S during cycling,thus inhibiting the LiPSs shuttling and boosting the redox reaction kinetics via a chemisorption-catalytic conversion mechanism.The enhanced wettability of the p-Fe_(3)O_(4)-NSs-based separator with the electrolyte enables fast transportation of lithium ions.Benefitting from these alluring properties,the functionalized separator with p-Fe_(3)O_(4)-NSs endows the battery with an admirable rate performance of 877 mAh g^(−1)at 2 C,an ultra-durable cycling performance of up to 2176 cycles at 1 C,and a promising areal capacity of 4.55 mAh cm^(−2)under high-sulfur-loading and lean-electrolyte conditions(4.29 mg cm^(−2),electrolyte/ratio:8μl mg^(−1)).This study will offer fresh insights on the rational design and low-cost fabrication of multifunctional separator to strengthen electrochemical reaction kinetics by regulating LiPSs conversion for developing efficient and long-life Li-S batteries.

Recovering metals from printed circuit board scrap by a mechanical separation process

A mechanical separation process was developed for recovering metals from printed circuit board(PCB) scrap;it included three steps:impact crushing,sieving and fluidization separation.The mechanism of the technique was based on the difference in the crushabilities of metallic and nonmetallic materials in the PCBs that led to the concentrated distribution of metals in particles of larger sizes and nonmetals mostly in particles of smaller sizes.It was found that crushed PCB particles from 0.125 mm to 1.000 mm contained about 80% of metals in the PCBs.Metals acquired satisfactory liberation in particles smaller than 0.800 mm.The crushed PCB particles were sieved into fractions of different size ranges.Each fraction separately went through a gas-solid fluidized bed operating at a selected optimal gas velocity for the specific size range.Approximately 95% of metals in printed circuit board particles from 0.125 mm to 0.800 mm was recovered by the gas-fluidized bed separator at the selected optimal gas velocity.However,separation of metals from particles smaller than 0.125 mm was not satisfactory.Further study is needed on metal recovery from fine particles.