Estimation of Density as a Function of Temperature and Pressure for Imidazolium-Based Ionic Liquids Using a Multilayer Net with Particle Swarm Optimization

The liquid density of imidazolium-based ionic liquids has been estimated using a combined method that includes an artificial neural network and a simple group contribution method. A total of 1736 data points of density at several temperatures and pressures, corresponding to 131 ionic liquids, have been used to train the neural network developed with particle swarm optimization. To discriminate among the different substances, the molar mass and the structure of the molecule were given as input variables. Then, new values of density as a function of temperature and pressure for 33 other ionic liquids (426 data points) have been predicted and the results compared to experimental data from the literature. The results show that the chosen artificial neural network with particle swarm optimization and the group contribution method represent an excellent alternative for the estimation of the liquid density of imidazolium-based ionic liquids with acceptable accuracy (AARD=0.44; R ² = 0.9934), for a wide range of temperatures and pressures (258 K to 393K and 99kPa to 206,940kPa).     

Symbolic regression via genetic programming in the optimization of a controlled release pharmaceutical formulation

Symbolic regression via genetic programming (GP) was used in the optimization of a pharmaceutical zero-order release matrix tablet, and its predictive performance was compared to that of artificial neural network (ANN) models. Two types of GP algorithms were employed: 1) standard GP, where a single population is used with a restricted or an extended function set, and 2) multi-population (island model) GP, where a finite number of populations is adopted. The amounts of four polymers, namely PEG4000, PVP K30, HPMC K100 and HPMC E50LV were selected as independent variables, while the percentage of nimodipine released in 2 and 8 h (Y_2h, and Y_8h), respectively, and the time at which 90% of the drug was dissolved (t_90%), were selected as responses. Optimal models were selected by minimization of the Euclidian distance between predicted and optimum release parameters. It was found that the prediction ability of GP on an external validation set was higher compared to that of the ANNs, with the multi population and standard GP combined with an extended function set, showing slightly better predictive performance. Similarity factor (f₂) values confirmed GP's increased prediction performance for multi-population GP (f₂ = 85.52) and standard GP using an extended function set (f₂ = 84.47).     

Prediction of mechanical properties of a warm compacted molybdenum prealloy using artificial neural network and adaptive neuro-fuzzy models

Predictive models using artificial neural network (ANN) and adaptive neuro-fuzzy inference system (ANFIS) were successfully developed to predict yield strength and ultimate tensile strength of warm compacted 0.85 wt.% molybdenum prealloy samples. To construct these models, 48 different experimental data were gathered from the literature. A portion of the data set was randomly chosen to train both ANN with back propagation (BP) learning algorithm and ANFIS model with Gaussian membership function and the rest was implemented to verify the performance of the trained network against the unseen data. The generalization capability of the networks was also evaluated by applying new input data within the domain covered by the training pattern. To compare the obtained results, coefficient of determination (R²), root mean squared error (RMSE) and average absolute error (AAE) indexes were chosen and calculated for both of the models. The results showed that artificial neural network and adaptive neuro-fuzzy system were both potentially strong for prediction of the mechanical properties of warm compacted 0.85 wt.% molybdenum prealloy; however, the proposed ANFIS showed better performance than the ANN model. Also, the ANFIS model was subjected to a sensitivity analysis to find the significant inputs affecting mechanical properties of the samples.     

Electronic expansion valve mass flow rate prediction based on dimensionless correlation and ANN model

Refrigerant mass flow rate through electronic expansion valve (EEV) makes significant sense for refrigeration system intelligent control and energy conservation. Objectives of this study were to present experimental data of R134a mass flow rate through EEV and to develop models for EEV mass flow rate prediction via two approaches: dimensionless correlation based on Buckingham π-theorem and artificial neural network (ANN) model based on dimensionless parameters. The database utilized for model training and test was comprised of our experimental data and data available in open literatures including R22, R407C, R410A and R134a. Compared with three existing dimensionless correlations, the proposed dimensionless correlation and ANN model demonstrated higher accuracy. The proposed dimensionless correlation gave mean relative error (MRE) of 6.60%, relative mean square error of (RMSE) 12.05 kg h⁻¹ and correlation coefficient (R²) of 0.9810. The ANN model with the configuration of 8-6-1 showed MRE, RMSE and R² of 3.97%, 7.59 kg h⁻¹ and 0.9924, respectively.     

Numerical simulation with finite element and artificial neural network of ball indentation for mechanical property estimation

A combined mechanical property evaluation methodology with ABI (Automated Ball Indentation) simulation and Artificial Neural Network (ANN) analysis is evolved to evaluate the mechanical properties for Carbon Manganese Steel (SA-333 Grade-6) and Stainless Steel (SS-304LN). The experimental load deflection data is converted into meaningful mechanical properties for these materials and their evaluated property is verified with experimental tensile specimen results. An ANN database is generated with the help of contact type finite element analysis by numerically simulating the ABI process for various magnitudes of yield strength (σ _yp ) (200 MPa–400 MPa) with a range of strain hardening exponent (n) (0.05–0.5) and strength coefficient (K) (600 MPa–1600 MPa). For the present problem, a ball indenter of 1.57 mm diameter having Young’s modulus higher than test piece is used to minimize the error due to indenter deformation. Test piece dimension is kept large enough in comparison to the indenter configuration in the simulation to minimize the deflection at the outer edge of the test piece. Further, this database after the neural network training; is used to analyse measured material properties of different test pieces. The ANN predictions are reconfirmed with contact type finite element analysis for an arbitrary selected test sample. The methodology evolved in this work can be extended to predict material properties for any irradiated nuclear material in the service. Extensions of the ABI tests and the associated database analysis could lead to evaluation of the indentation energy to fracture needed for the structural integrity assessment of aged components.     

Supramolecular host–guest systems constructed of pyrrole derivatives and 3D-coordination polymers

Pyrrole derivatives have been shown to be completely or partially oxidized within the expandable channels of the 3D-coordination polymers [(R₃Sn)₃Fe(CN)₆] _n and [(R₃Sn)(R₃^′Sn)₂Fe(CN)₆] _n , R and R′ = Me, n-Bu, or Ph, to give novel class of supramolecular host–guest systems. The structures and physical properties of these host–guest systems depend on the reaction time, nature of the host and guest, the space empty within the network of the 3D-coordination polymers. Pyrrole undergoes oxidative polymerization in the channels of the 3D-coordination polymers to form semiconducting diamagnetic supramolecular host–guest systems. Whereas N-methylpyrrole and 2,5-dimethylpyrrole are not polymerized under these experimental conditions, but give paramagnetic charge transfer (CT) supramolecular host–guest systems.     

Comparative study between ANN models and conventional equations in the analysis of fatigue failure of GFRP

The purpose of this paper is to assess the applicability of two artificial neural networks (ANN) architecture, perceptron ANN, modular ANN, and Adam’s equation in the modeling of fatigue failure in polymer composites, more specifically in glass fiber reinforced plastic (GFRP). In the application of the model using ANN we show the feasibility of obtaining good results with a small number of S–N curves. The other model used involves applying empirical equations known as Adam’s equations. A comparative study on the application of the aforementioned models is developed based on statistical tools such as correlation coefficient and mean square error. For this analysis we used composite materials in the form of laminar structures with distinct stacking sequences, which are applied industrially in the construction of large reservoirs. Reinforcements consist of mats and bidirectional textile fabric made of E-glass fibers soaked in unsaturated orthophthalic polyester resin. These were tested for six different stress ratios: R = 1.43, 10, ?1.57, ?1, 0.1, and 0.7. The results showed that although ANN modeling is in the initial phase, it has great application potential.     

Detailed Physical Characterizations of Cu-rich and Ru-poor (Ru0.9Cu0.1)Sr2YCu2O7.9

The optimized nominal composition, (Ru_0.9Cu_0.1) Sr₂YCu₂O_7.9 sample, has been prepared through high-pressure and high-temperature solid-state densification method. The obtained material has been studied by X-ray (laboratory) diffraction powder technique, magnetization and detailed magneto-transport measurements. The title compound indicates bulk magneto-superconducting properties under field strengths of H=10, 100, 500 and 1000 Oe. It shows diamagnetic transition at T _d=54, 38, 20 and 8 K for H=10, 100, 500 and 1000 Oe, respectively, in the zero-field-cooled susceptibility measurements. The high-field (H=5 and 10 kOe) molar susceptibility measurements show sharp ferromagnetic transition at ∼150 K with reduced molar susceptibility values. The various field dependence of magnetization, M(H), isotherm curves recorded at constant temperatures (5, 10, 25, 50, 100 and 150 K) indicate ferromagnetic saturation, whereas the MH curves measured at 200 and 300 K conditions reveal the paramagnetic state of the compound. Though the sample showed onset transition temperature, T_c^onsetT_{\mathrm{c}}^{\mathrm{onset}}, at ∼34 K under different field strengths (H=0, 10, 30, 50, 70 and 90 kOe), no T_c^R=0T_{\mathrm{c}}^{R=0} is seen down to 2 K. Even under relatively low applied field (ΔH=10 kOe) the title compound shows large negative magnetoresistance (MR) of about 68% at 2 K and increases with increasing the field strength up to ΔH=90 kOe (MR=77% at 2 K). This value is amazing and probably higher than other 1212 type ruthenocuprates. The title compound which shows little negative MR (about 1%) in the high temperature regions (125–300 K) is not affected much by different field strengths. Among the different fixed temperature MR(H) isotherms, the MR(H) curve measured at 5 K shows maximum negative MR of about 47% at 90 kOe compared to other four (T=50, 100, 200 and 300 K) MR(H) curves.     

Application of artificial neural network for prediction of the oxidation behavior of aluminized nano-crystalline nickel

In this paper, the applicability of artificial neural network (ANN) for the prediction of the oxidation kinetics of aluminized coating is presented. For developing the model, a consistent set of experimental data i.e. nanocrystalline Ni samples were aluminized by two steps aluminizing process and oxidized at 800, 900 and 1000 °C for various times are used. The exposure time and temperature of oxidation were used as the inputs of the model and the resulting mass gain of oxidized samples as the output of the model. Multi-layer perceptron neural network structure and back-propagation algorithm are used for the training of the model. After testing many different ANN architectures an optimal structure of the model i.e. 2-5-6-1 is obtained. Comparison of experimental and predicted values using the proposed ANN model shows that there is a good agreement between them with mean relative error less than 1.2%. This shows that the ANN model is an accurate and reliable approach to predict the oxidation behavior of aluminized nanocrystalline coatings.     

Integrated ANN–GA for estimating the minimum value for machining performance

In this study, we proposed a new approach in estimating a minimum value of machining performance. In this approach, artificial neural network (ANN) and genetic algorithm (GA) techniques were integrated in order to search for a set of optimal cutting condition points that leads to the minimum value of machining performance. Three machining cutting conditions for end milling operation that were considered in this study are speed (v), feed (f) and radial rake angle (γ). The considered machining performance is surface roughness (R _a). The minimum R _a value at the optimal v, f and γ points was expected from this approach. Using the proposed approach, named integrated ANN–GA, this study has proven that R _a can be estimated to be 0.139?µm, at the optimal cutting conditions of f?=?167.029?m/min, v?=?0.025?mm/tooth and γ?=?14.769°. Consequently, the ANN–GA integration system has reduced the R _a value at about 26.8%, 25.7%, 26.1% and 49.8%, compared to the experimental, regression, ANN and response surface method results, respectively. Compared to the conventional GA result, it was also found that integrated ANN–GA reduced the mean R _a value and the number of iterations in searching for the optimal result at about 0.61% and 23.9%, respectively.     

基于神经网络的电弧增材制造铝合金力学性能预测

目的 预测不同工艺参数下电弧增材制造铝合金的力学性能。方法 通过实验建立了电弧增材制造6061铝合金及Ti C增强6061铝合金力学性能的数据集,并建立了一种以焊接电流、焊接速度、脉冲频率、TiC颗粒含量为输入,以屈服强度和抗拉强度为输出的神经网预测模型,对比了反向传播神经网络(BP)、粒子群算法优化BP神经网络(PSO-BP)、遗传算法优化BP神经网络(GA-BP)3种预测模型的精度。结果 与BP模型和PSO-BP模型相比,GA-BP预测模型具有更好的预测精度。其中,GA-BP模型预测6061铝合金屈服强度最佳结果的相关系数(R)为0.965,决定系数(R²)为0.93,平均绝对误差(Mean Absolute Error,MAE)为2.35,均方根误差(Root Mean Square Error,RMSE)为2.67;预测Ti C增强的6061铝合金抗拉强度最佳结果的R=1,R²高达0.99,MAE为0.46,RMSE为0.49,GA-BP具有良好的预测精度。结论 BP、PSO-BP、GA-BP 3种神经网络模型可以用来预测电弧增材制造...     

Ultraflat indium tin oxide films prepared by ion beam sputtering

Indium tin oxide (ITO) films with a smooth surface (root-mean-square roughness; R_rms=0.40 nm) were made using a combination of the deposition conditions in the ion beam-sputtering method. Sheet resistance was 13.8 Ω/sq for a 150-nm-thick film grown at 150 °C. Oxygen was fed into the growth chamber during film growth up to 15 nm, after which, the oxygen was turned off throughout the rest of the deposition. The surface of the films became smooth with the addition of ambient oxygen but electrical resistance increased. In films grown at 150 °C with no oxygen present, a rough surface (R_rms=2.1 nm) and low sheet resistance (14.4 Ω/sq) were observed. A flat surface (R_rms=0.5 nm) with high sheet resistance (41 Ω/sq) was obtained in the films grown with ambient oxygen throughout the film growth. Surface morphology and microstructure of the films were determined by the deposition conditions at the beginning of the growth. Therefore, fabrication of ITO films with a smooth surface and high electrical conductivity was possible by combining experimental conditions.     

Densification of (Na,K)NbO<Subscript>3</Subscript> piezoelectric ceramics by two-step mixing process

Lead-free Na_0.5K_0.5NbO₃ (NKN) piezoelectric ceramics were sintered with a new process, “two-step mixing process,” in which a part of alkali source powders was initially preserved and mixed with the rest matrix powders after the calcinations step. As a result, the sintering of NKN ceramics was improved, and the sample sintered at 1082 °C with the initial preservation ratio (R _A) of 5% demonstrated the highest density of ρ = 4.38 g/cm³ (97.1% of the theoretical density), compared with ρ = 4.36 g/cm³ (96.7% of the theoretical density) for the non-preservation specimen (R _A = 0%). The former sample showed the best piezoelectric constant of d ₃₃ = 125 pC/N and electromechanical coupling coefficient of k _p = 0.42, while the latter sample had d ₃₃ = 116 pC/N and k _p = 0.37. These results indicated that the two-step mixing process was effective for the sintering of lead-free NKN ceramics, despite no sintering additive and cold isostatic pressing were used.     

Application of neural networks for surface roughness measurement in finish turning

A laser system which incorporates a charge-coupled-device (CCD) sensor is developed to measure, in real lime, the maximum peak-to-valley surface roughness, R _max. produced during finish turning. A three-layer neural network is used in conjunction with a back-propagation learning algorithm to predict R _max, by quickly recognizing the angular scattered light patterns (ASLP) reflected from the workpiece in the feed direction. The predicted R _max values have a maximum error of about 10% when compared to conventional stylus measurements.     

The investigation of the effect of particle size on wear performance of AA7075/Al2O3 composites using statistical analysis and different machine learning methods

In this study, 7075 - Al₂O₃ (5 wt%) composites with a particle size of 0.3 µm, 2 µm, and 15 µm were developed by hot pressing. The dry sliding wear performance of the specimens was evaluated under loads of 5 N, 10 N, 20 N, 30 N, and at sliding speeds of 80 mm/s, 110 mm/s, 140 mm/s by reciprocating wear tests. The wear tests showed that 7075 - 5Al₂O₃ (15 µm) exhibited the best wear performance. The volume loss of 7075 - 5Al₂O₃ (15 µm) under load of 30 N for sliding speed of 140 mm/s was 37.1% lower than the unreinforced 7075 alloy. The volume loss (mm³) of composites reinforced with the particle size of 0.3 µm, 2 µm, and 15 µm was 11.62, 9.87, and 8.07, respectively, for load of 30 N and sliding speed of 140 mm/s. An increase in the applied load and sliding speed increased the wear severity by changing the wear mechanism from abrasion to delamination. The analysis of variance (ANOVA) showed that the load was the most significant parameter on the volume loss. The linear regression (LR), support vector regression (SVR), artificial neural network (ANN), and extreme learning machine (ELM) were used for the prediction of volume loss. The determination coefficient (R²) of the LR, SVR, ANN, and ELM was 0.814, 0.976, 0.935, and 0.989, respectively. The ELM model has the highest success. Thus, the ELM model has significant potential for the prediction of wear behaviour for Al matrix composites.     

Single crystal structure determination and infrared fluorescence of the system (K3Sr1−xNdx) (Nd1−xSr1+x) Nb10O30

The μ-pulling down technique has permitted to grow single crystal fibers, of the gross chemical formula K₃Sr₂NdNb₁₀O₃₀, having a sufficient optical quality to carry out spectroscopic studies. The crystal structure has been determined and refined to reliability factors: (i) R₁ = 0.0384 (wR₂ = 0.0665) at room temperature; (ii) R₁ = 0.0334 (wR₂ = 0.0638) at 120 K. Difference in the cationic distribution over the 15- and 12-fold sites was noticed. IR fluorescence spectra investigated under different laser excitation wavelength at 300 K and 77 K show strong emissions at 0.9 and 1.06 μm. Low temperature fluorescence behavior is compatible with Nd³⁺ ions located in both Sr²⁺ and K⁺ sites with 15- and 12-fold coordination, respectively.     

Preparation of Li–Al–O films by laser chemical vapor deposition

Li–Al–O films were prepared on AlN substrates by laser chemical vapor deposition at deposition temperatures (T_dep) of 800–1300 K and molar ratios of Li to Al precursors (R_Li/Al) of 0.1–12. Single-phase α-LiAl₅O₈ films having faceted grains with pyramidal and polygonal shapes were obtained at T_dep = 1107–1280 K and R_Li/Al = 0.1–2.9. Single-phase γ-LiAlO₂ films having pyramidal grains were prepared at T_dep = 984–1238 K and R_Li/Al = 0.9–10.6. Under the conditions of T_dep = 923 K and R_Li/Al = 11.4, single-phase β-Li₅AlO₄ films with a fluffy morphology were deposited. The highest deposition rate of Li–Al–O films was 98 μm h⁻¹ with a mixture of γ-LiAlO₂ and β-Li₅AlO₄ at T_dep = 944 K.     

Modelling of a thermal insulation system based on the coldest temperature conditions by using artificial neural networks to determine performance of building for wall types in Turkey

In formation of building external envelope, as two important criteria, climatic data and wall types must be taken into consideration. In the selection of wall type, the thickness of thermal insulation layer (d_i) must be calculated. As a new approach, this study proposes determining the thermal insulation layer by using artificial neural network (ANN) technique. In this technique five different wall types in four different climatic regions in Turkey have been selected. The ANN was trained and tested by using MATLAB toolbox on a personal computer. As ANN input parameters, U_w, T_e,Met, T_e,TSE, R_wt, and q_TSE were used, while d_i was the output parameter. It was found that the maximum mean absolute percentage error (MRE, %) is less than 7.658%. R² (%) for the training data were found ranging about from 99.68 to 99.98 and R² for the testing data varied between 97.55 and 99.96. These results show that ANN model can be used as a reliable modeling method of d_i studies.     

Magnetic and Dielectric Properties of <Emphasis Type="Italic">R</Emphasis><Subscript>2</Subscript>CuTiO<Subscript>6</Subscript> Compounds (<Emphasis Type="Italic">R</Emphasis>=Y,La, Pr and Nd)

Magnetic and dielectric properties of the double perovskite compounds of the type R ₂CuTiO₆ (RCTO, where R=Y, La, Pr and Nd) has been studied. Y₂CuTiO₆ (YCTO) crystallizes in a hexagonal unit cell, whereas the other three compounds form into orthorhombic structure. All four compounds show paramagnetic behavior down to 5 K. The dielectric studies show moderate dielectric constant (ε′) and very small dielectric loss (tan δ) for YCTO. The orthorhombic members of RCTO compounds exhibit moderate values of ε′ and tan δ. The dielectric properties are presented and discussed here in the light of the influence of structure and rare-earth ions on the physical properties of RCTO compounds.