Object-based hyperspectral image classification using a new latent block model based on hidden Markov random fields

Pattern Analysis and Applications - Efficient high-dimensional analyses of hyperspectral datasets and their utilization within classification algorithms is a popular topic in the field of data...     

A review of machine learning applications in IoT-integrated modern power systems

This paper aims to provide a systematic and comprehensive survey of state-of-the-art machine learning techniques and their potential applications in IoT-integrated power systems.     

Experimental investigation on the mean drop size and drop size distribution in an L-shaped pulsed sieve-plate column

Vertical pulsed extraction columns cannot be employed in applications with height limitations. On the other hand, the horizontal extraction columns have low throughput, which affects their applicability in industrial applications. Therefore, there is a need to design a new type of extractors for such circumstances. In this paper, an experimental study on drop sizes has been implemented in a novel L-shaped pulsed sieve-plate extraction column in the absence and presence of mass transfer. Moreover, new correlations are developed for prediction of the mean drop size and size distribution using the log-normal probability density function.     

Core–shell structured Cu@Pt nanoparticles as effective electrocatalyst for ethanol oxidation in alkaline medium

Core–shell nanoparticles of Cu@Pt/C electrocatalysts were synthesized using various Pt:Cu atomic ratios with NaBH₄ as a reducing agent. The crystal structure and surface morphology were studied using X-ray diffraction (XRD) and transmission electron microscopy (TEM). The incorporation of copper in Cu@Pt/C electrocatalysts was found to shift all Pt diffraction planes in the negative direction with expanding the crystal lattice dimensions. The electrocatalytic activity of various Cu@Pt/C electrocatalysts was investigated using cyclic voltammetry, chronoamperometry and electrochemical impedance spectroscopy. Cu@Pt/C electrocatalysts containing Pt:Cu atomic ratios of 1:0.5 and 1:2 showed an enhanced electrochemical performance for ethanol oxidation when related to Pt/C.     

Novel composite membrane based on zirconium phosphate-ionic liquids for high temperature PEM fuel cells

Composite membranes composed of zirconium phosphate (ZrP) and imidazolium-based ionic liquids (IL), supported on polytetrafluoroethylene (PTFE) were prepared and evaluated for their application in proton exchange membrane fuel cells (PEM) operating at 200 °C. The experimental results reported here demonstrate that the synthesized membrane has a high proton conductivity of 0.07 S cm^?1, i.e, 70% of that reported for Nafion. Furthermore, the composite membranes possess a very high proton conductivity of 0.06 S cm^?1 when processed at 200 °C under completely anhydrous conditions. Scanning electron microscopy (SEM) images indicate the formation of very small particles, with diameters in the range of 100–300 nm, within the confined pores of PTFE. Thermogravimetric analysis (TGA) reveals a maximum of 20% weight loss up to 500 °C for the synthesized membrane. The increase in proton conductivity is attributed to the creation of multiple proton conducting paths within the membrane matrix. The IL component is acting as a proton bridge. Therefore, these membranes have potential for use in PEM fuel cells operating at temperatures around 200 °C.     

A new manufacturable filter design approach for spectral reflectance estimation

Selection of the best possible filter set among a set of available filters is the obvious method of increasing dimension of camera signals for spectral reflectance reconstruction. There are also methods that are focusing on the filter design regardless of noticing to the constructability of the designed filters. This study shows that direct optimization of physical variables of filter manufacturing technique is more reliable than indirect approach of designing and then physical manufacturing of the designed filters. Direct optimization of the transmission‐controlling primaries in filter manufacturing process would guarantee having the designed filters in reality. Combination of some solvent dyes was used as the spectral transmission matching system for filter manufacturing. As a conventional technique, filters were designed and best possible dye concentrations that match the desired filters were calculated. As an alternative approach, filters were also designed using direct optimization of dyes concentrations. The results showed that direct optimization of dye concentrations exhibits better performance in comparison with the conventional technique. © 2016 Wiley Periodicals, Inc. Col Res Appl, 42, 316–326, 2017     

Simultaneous multi-objective optimization of a new promoted ethylene dimerization catalyst using grey relational analysis and entropy measurement

A hybrid approach between the Taguchi method and grey relational analysis (GRA) with entropy measurement was applied to determine a single optimum setting for reaction factors of the proposed ethylene dimerization catalyst having overall selectivity to 1-butene (S_1-btn (%)) and turnover frequency (TOF (h^-1)) as multiple quality characteristics. Titanium tetrabutoxide (Ti(OC₄H₉)₄) catalyst precursor in combination with triethyl aluminum (TEA) activator, 1,4-dioxane as a suitable modifier, and ethylene dichloride (EDC) as a novel promoter were used in the catalysis. Control factors of temperature, pressure, Al/Ti, 1,4-dioxane/Ti, and EDC/Ti mol ratios were investigated on three levels and their main effects were discussed. The effect of the binary interaction between temperature, pressure, and Al/Ti mol ratio was also examined. Weight of the responses was determined using entropy. Analysis of variance (ANOVA) for data obtained from GRA indicated that EDC/Ti mol ratio with 27.64% contribution had the most profound effect on the multiple quality characteristics. Development of the weighted Grey-Taguchi method used the Taguchi method as its basic structure, adopted GRA to deal with multiple responses, and entropy to enhance the reasonability of the comprehensive index produced by GRA to make the results more objective and accurate. Overall, these combined mathematical techniques improved catalytic performance for 1-butene production.     

On estimating the interface normal and curvature in piecewise linear interface calculation-volume of fluid approach for three-dimensional arbitrary meshes

Volume of fluid (VOF) method with its Piecewise Linear Interface Calculation (PLIC) reconstruction algorithm is one of the most popular approaches in numerical simulation of interfacial flows with a wide range of applications in different areas. In an effort to evaluate the similarity of the PLIC-generated planes in comparison with the exact interface, a point-cloud, based on the polygon centers of PLIC planes is extracted, which later is used to form a triangular grid that represents the estimated interface. The main objective of this article is to evaluate the interface geometrical properties based on the extracted triangular grid of the interface. The methods presented in this article, characterized by a higher spatially convergence ratio, are compared with the commonly used methods. The proposed methods are tested for two 3-dimensional general test cases, where an evident improvement is seen in calculation accuracy and spatial convergence of the errors of interface normal vector and curvature.     

Concentration dependence of viscometric properties of model short chain polymer solutions

We study the concentration dependence of the conformational and viscometric behaviour of short-chain polymer solutions in shear flow by conducting a series of non-equilibrium molecular dynamics simulations, covering the entire concentration range. Our model explicitly incorporates all of the important generic features of real polymer solutions—excluded volume, hydrodynamic interactions and finite chain extensibility. Hydrodynamic interactions are included exactly by treating the solvent explicitly as an atomic fluid. The polymer molecules studied consist of 20-site bead-rod model molecules, which correspond approximately to 12 Kuhn steps in the melt. For polyethylene, this represents a molar mass of 1800 g mol⁻¹. In some respects, our results are consistent with experimental and theoretical results obtained for long-chain polymer solutions. We calculate the Flory-Fox constant and find a value that agrees reasonably well with results for long chain polymer solutions. Due to the short chain length of the molecules investigated, no semidilute region exists for these solutions. However, the radius of gyration and viscosity still exhibit strong concentration dependence, which is well described by power series, rather than power law expressions, in contrast to the behaviour usually observed in long-chain polymer solutions.     

Well-defined core-shell nanoparticles containing cyclodextrin in the shell: A comprehensive study

The purpose of this study was to conceive novel nanoparticles (NPs) designed for drug delivery applications. The NPs were based on a hydrophobic poly (lactic acid) (PLA) core and a hydrophilic β-cyclodextrin polymer (Poly-β-CD) shell. The PLA NPs were prepared by the nanoprecipitation method. The Poly-β-CD shell was obtained by adsorption from Poly-β-CD solution. Different complementary techniques have been used to fully characterize the system including dynamic light scattering, Zeta potential measurements, solution proton nuclear magnetic resonance spectroscopy, transmission electron microscopy, and small angle neutron scattering. They provide good evidence of a core-shell structure. The NPs parameters are dependent on the elaboration process, Poly-β-CD characteristics and adsorption conditions. In the conditions of adsorption’ saturation, the shells were constituted of Poly-β-CD monolayers covering the PLA surfaces with adsorption amounts from 2 to 4 mg/m², allowing reaching Poly-β-CD weight fractions of 10-20% in the NPs.