A comparison of machine learning methods for cutting parameters prediction in high speed turning process

Support vector machines are arguably one of the most successful methods for data classification, but when using them in regression problems, literature suggests that their performance is no longer state-of-the-art. This paper compares performances of three machine learning methods for the prediction of independent output cutting parameters in a high speed turning process. Observed parameters were the surface roughness (Ra), cutting force \((F_{c})\), and tool lifetime (T). For the modelling, support vector regression (SVR), polynomial (quadratic) regression, and artificial neural network (ANN) were used. In this research, polynomial regression has outperformed SVR and ANN in the case of \(F_{c}\) and Ra prediction, while ANN had the best performance in the case of T, but also the worst performance in the case of \(F_{c}\) and Ra. The study has also shown that in SVR, the polynomial kernel has outperformed linear kernel and RBF kernel. In addition, there was no significant difference in performance between SVR and polynomial regression for prediction of all three output machining parameters.     

The effect of fusing Sentinel-2 bands on land-cover classification

The Sentinel-2 satellite currently provides freely available multispectral bands at relatively high spatial resolution but does not acquire the panchromatic band. To improve the resolution of 20 m bands to 10 m, existing pansharpening methods (Brovey transform [BT], intensity–hue–saturation [IHS], principal component analysis [PCA], the variational method [P + XS], and the wavelet method) required adjustment, which was achieved using higher resolution multispectral bands in the role of a panchromatic band to fuse bands at a lower spatial resolution. After preprocessing, six bands at lower resolution were divided into two groups because some image fusion methods (e.g. BT, IHS) are limited to a maximum of three input bands of a lower resolution at a time. With respect to the spectral range, the higher resolution band for the first group was synthesized from bands 4 and 8, and band 8 was selected for the second group. Given that one of the main remote sensing applications is land-cover classification, the classification accuracy of the fusion methods was assessed as well as the comparison with reference bands and pixels. The supervised classification methods were Maximum Likelihood Classifier, artificial neural networks, and object-based image analysis. The classification scheme contained five classes: water, built-up, bare soil, low vegetation, and forest. The results showed that most of the fusion methods, particularly P + XS and PCA, improved the overall classification accuracy, especially for the classes of forest, low vegetation, and bare soil and in the detection of coastlines. The least satisfying results were obtained from the wavelet method.     

Synthesis and characterization of organic–inorganic hybrids based on epoxy resin and 3‐glycidyloxypropyltrimethoxysilane

Organic–inorganic hybrid materials based on diglycidyl ether of bisphenol A (DGEBA) and 3‐glycidyloxypropyltrimethoxysilane (GLYMO) were prepared, using a poly(oxypropylene) diamine Jeffamine D230 as a curing agent. Materials were studied by differential scanning calorimetry (DSC), thermogravimetric analysis (TGA), infrared spectroscopy, swelling in tetrahydrofurane (THF), and Soxhlet extraction in THF. A dependence of the final conversion of epoxy groups and the final degree of organic phase crosslinking on inorganic phase content was found. The inorganic phase presents a steric hindrance to full crosslinking of epoxy groups. It also immobilizes the organic chains and improves the temperature stability of hybrid materials. Products of GLYMO hydrolysis together with unreacted organic molecules lower the glass transition temperature of hybrid materials. © 2004 Wiley Periodicals, Inc. J Appl Polym Sci 92: 498–505, 2004     

State equation of mineral sands

The pressure-volume analogy between compressible fluids and macroscopic sand bodies (Ivsic et al. in Phys A, 277:47–61, 2000) is further extended using quantitative determination of corresponding empirical constants based on adapted van der Waals state equation. The isothermal constants obtained by interpretation of triaxial sand tests at so called “critical state of sand” are clearly related to the universal ideal gas properties and molar properties of mineral sands. The corresponding constants for sand and gases or other volatile liquids have the same order of magnitude. The apparent bulk repulsion/attraction effects in sand bodies are also discussed.     

Hydroxyapatite formation from cuttlefish bones: kinetics

Highly porous hydroxyapatite (Ca₁₀(PO₄)₆·(OH)₂, HA) was prepared through hydrothermal transformation of aragonitic cuttlefish bones (Sepia officinalis L. Adriatic Sea) in the temperature range from 140 to 220°C for 20 min to 48 h. The phase composition of converted hydroxyapatite was examined by quantitative X-ray diffraction (XRD) using Rietveld structure refinement and Fourier transform infrared spectroscopy (FTIR). Johnson–Mehl–Avrami (JMA) approach was used to follow the kinetics and mechanism of transformation. Diffusion controlled one dimensional growth of HA, predominantly along the a-axis, could be defined. FTIR spectroscopy determined B-type substitutions of CO₃ ²⁻ groups. The morphology and microstructure of converted HA was examined by scanning electron microscopy. The general architecture of cuttlefish bones was preserved after hydrothermal treatment and the cuttlefish bones retained its form with the same channel size (~80 × 300 μm). The formation of dandelion-like HA spheres with diameter from 3 to 8 μm were observed on the surface of lamellae, which further transformed into various radially oriented nanoplates and nanorods with an average diameter of about 200–300 nm and an average length of about 8–10 μm.     

Mixed meshless formulation for analysis of shell-like structures

An efficient mixed meshless computational method based on the Local Petrov–Galerkin approach for analysis of plate and shell structures is presented. A concept of a three-dimensional solid is applied allowing the use of complete three-dimensional constitutive equations, and exact shell geometry can be described. Discretization is carried out by using both the moving least square approximation and the polynomial functions. Independent field variables are the strain and stress tensor components expressed in terms of the nodal values, which are then replaced by the nodal displacements by using the independent displacement interpolation. A closed global system of equations with only nodal displacements as unknown variables is derived. The undesired locking phenomena are fully suppressed. The proposed mixed formulation is numerically more efficient than the available meshless fully displacement approach, as demonstrated by the numerical examples.     

Revisiting the Ziegler–Nichols process dynamics characterization

The Ziegler–Nichols process dynamics characterization is based on the estimation of the ultimate gain k_u and ultimate frequency ω_u. The angle φ of the tangent to the Nyquist curve at the frequency ω_u is introduced, as an additional parameter in the frequency domain. The essential dynamic characteristics of the process can be captured by using the tangent rule, proposed here as an extension of the Ziegler–Nichols approach. The validity of the tangent rule is confirmed by using the PID controller optimization on a test batch consisting of stable, integrating and unstable processes, including dead-time. Parameters k_u, ω_u and φ can be determined from the sustained oscillations, using the phase-locked loop identifier module.     

Immunity of electronic and transport properties of phosphorene nanoribbons to edge defects

We present an extensive study of the electronic properties and carrier transport in phosphorene nanoribbons (PNRs) with edge defects by using rigorous atomistic quantum transport simulations. This study reports on the size- and defect-dependent scaling laws governing the transport gap, and the mean free path and carrier mobility in the PNRs of interest for future nanoelectronics applications. Our results indicate that PNRs with armchair edges (aPNRs) are more immune to defects than zig-zag PNRs (zPNRs), while both PNR types exhibit superior immunity to defects relative to graphene nanoribbons (GNRs). An investigation of the mean free path demonstrated that even in the case of a low defect density the transport in PNRs is diffusive, and the carrier mobility remains a meaningful transport parameter even in ultra-small PNRs. We found that the electron–hole mobility asymmetry (present in large-area phosphorene) is retained only in zPNRs for W > 4 nm, while in other cases the asymmetry is smoothed out by edge defect scattering. Furthermore, we showed that aPNRs outperform both zPNRs and GNRs in terms of carrier mobility, and that PNRs generally offer a superior mobility-bandgap trade-off, relative to GNRs and monolayer MoS₂. This work identifies PNRs as a promising material for the extremely scaled transistor channels in future post-silicon electronic technology, and presents a persuasive argument for experimental work on nanostructured phosphorene.

Open image in new window