Support-vector-regression-based prediction of water holdup in horizontal oil-water flow by using a bicircular conductance probe array

This paper presents a water holdup prediction method based on support vector regression (SVR) for horizontal oil-water two-phase flow when using a bicircular conductance probe array that consists of 24 conductance probes. The support vector machine (SVM) was employed to establish a nonlinear SVR model mapping the probe array responses into water holdup directly. Experiments were carried out in the 16 m long and 125 mm inner diameter horizontal pipe of an industrial scale experimental setup. The experimental data obtained under 220 flow conditions were first divided into modeling data set and comparing data set. The modeling data set is used for establishing a nonlinear SVR and a linear least squares regression (LSR) models, while the comparing data set is used for comparing both models with the equi-weight and optimal weight estimate methods. Comparison results obtained by using the comparing data set show that when the binary data of the probes’ responses are used only, the measurement accuracy of the optimal weight estimate method is the best. If the analog data can be obtained, the measurement accuracy of both regression methods are better than those of both weighting estimate methods, especially, the nonlinear SVR method provide the best measurement accuracy.     

Slug frequency in high viscosity oil-gas two-phase flow: Experiment and prediction

The number of slug units that traverses a particular point at a given time within a defined pipe cross-section is known as slug frequency. The behaviour of this critical parameter for two-phase flow in high viscosity oils is significantly different from those of conventional oils (of less than 1 Pa s). In this experimental investigation, new data on slugging frequency in high viscosity oil-gas flow are reported. Scaled experiments were carried out using a mixture of air and mineral oil in a 17 m long horizontal pipe of 0.0762 m ID. A high-speed Gamma Densitometer of frequency of 250 Hz was used for data acquisition over a time interval of 30 s. For the range of flow conditions investigated, increase in oil viscosity was observed to strongly influence the slug frequency. Comparison of the present data with prediction models available in the literature revealed discrepancies. A new correlation incorporating the effect of viscosity on slug frequency has been proposed for horizontal flow. The proposed correlation will improve the prediction of slug frequency in high viscosity oils.     

Estimation of biosurfactant yield produced by Klebseilla sp. FKOD36 bacteria using artificial neural network approach

Environmental and medium parameters estimation is an essential step in Bioprocess engineering. In the present study, artificial neural network (ANN) was employed in estimation of biosurfactants yield from bacterial strain Klebseilla sp. FKOD36, surface tension reduction as well emulsification index. The data obtained from experimental design were used in modelling and optimization of ANN method. Temperature, pH value, incubation period, carbon, nitrogen and hydrocarbon sources were used as input of ANN model independently in the prediction of biosurfactants yield, surface tension reduction and emulsification index. Using the optimized values of critical input elements of ANN, the experimental values of biosurfactant yield, emulsification index and surface tension showed close agreement with the model estimate. The most efficient ANN model assessment was 0.030 g/l for actual value 0.038 g/l of biosurfactant yield, 31.67% for actual value 31.68% of emulsification index, and 21.6 dyne/cm for actual value 21.5 dyne/cm of surface tension respectively.     

Multi input–single output models identification of tower bridge movements using GPS monitoring system

In this paper, RTK-GPS system was used for movement data collection. Two identification models namely; Multi input–single output (MISO) robust fit regression and Neural Network Auto-Regression Moving Average with eXogenous input (NNARMAX) models were used for the identification of these data. The analysis of test results indicate that: (1) the NNARMAX [4 4 1 1] and [5 4 1 5] models defined by taking into account the results of robust regression analysis estimate structural movements more accurately than the NNARMAX [0 1 0 0] model, and (2) the robust fit regression models have good capacities for mapping relationship of applied loads effects factors and displacements of tower. However, temperature and humidity effects on the entire modal shapes are insignificant and (3) the traffic loads are the main factor affects tower bridge displacement.     

Continuous accurate pH measurements of human GI tract using a digital pH-ISFET sensor inside a wireless capsule

This paper describes the design and implementation of a reliable, low-power pH-ISFET sensor inside a wireless diagnostic capsule for monitoring the pH of the human gastrointestinal (GI) tract. The sensitive element of the pH sensing device was an ion-sensitive FET (ISFET) with a threshold voltage that varies by concentration of hydrogen ions. An Ag/AgCl reference electrode that was surrounded by gel (mixed KCl) was applied in this pH-sensing device. An application-specific integrated circuit (ASIC) that is designed by a very large scale integration (VLSI) architecture was used as the interface between the pH sensing device and a wireless transceiver. Experimental results demonstrated that the pH sensitivity and operating range of the pH sensing device was 44.94 mV/pH and 1–11 pH; the device is linear. With the 18-bit Analog/Digital converter (ADC) module in the ASIC, the resolution could achieve 11,780 bits/pH. Moreover, the power dissipation of the pH-sensing device was only 0.048 mW while working in intermittent mode (duty cycle = 20%). Results from human experiments showed that this pH sensing device can work reliably for 136 h, and the pH data of sampled from the human GI tract can be received by a portable data recorder outside the human body in real time.     

Surface roughness prediction for the milling of Ti–6Al–4V ELI alloy with the use of statistical and soft computing techniques

This study focuses on Ti–6Al–4V ELI titanium alloy machining by means of plain peripheral down milling process and subsequent modeling of this process, in order to predict surface quality of the workpiece and identify optimal cutting parameters, that lead to minimum surface roughness. For the purpose of accomplishing this task a set of experiments were performed on a CNC milling centre and design of experiments based on Box Behnken Design (BBD) for a three factor and three level central composite design concept was conducted. Depth of cut, cutting speed and feed rate were selected as input parameters and surface roughness was measured after each experiment performed. At first, Response Surface Methodology (RSM) was employed for establishing a quadratic relationship between input and output parameters. Analysis of variance (ANOVA) was then conducted for the evaluation of the proposed formula. RSM was also used for the optimization analysis that followed for the determination of milling cutting parameters for minimum surface roughness. The analysis indicates that the use of BBD can reduce the number of experiments needed for modeling and optimizing the milling operation of Titanium alloys. Furthermore, this method is able to provide models that can reliably be used for any cutting conditions within the limits of the input data. Finally, Artificial Neural Networks (ANN) models were developed to allow for a more robust simulation model to be built and comparison between ANN and RSM models to be performed. From the presented results, for RSM, the mean square error and the correlation coefficient were determined to be 8.633 × 10⁻³ and 0.9713, respectively; for ANN models, the corresponding values were 2 × 10⁻³ and 0.9824, for the test group of the optimum model. Simulations indicated that, although input data were too few, a considerably reliable ANN model was able to be built and despite of its complexity compared to RSM model, it was proven to be superior in terms of prediction accuracy.     

Artificial neural network modeling of thin layer drying behavior of municipal sewage sludge

The back-propagation (BP) and generalized regression neural network models (GRNN) were investigated to predict the thin layer drying behavior in municipal sewage sludge during hot air forced convection. The accuracy of the BP model to predict the moisture content of the sewage sludge thin layer during hot air forced convective drying was far higher than that of the GRNN model. The GRNN models could automatically determine the best smoothing parameters, which were 0.6 and 0.3 for predicting the moisture content and average temperature, respectively. The model type for predicting the average temperature of the sewage sludge thin layer was selected for different sample groups by comparing their MSE values or R² values. The GRNN model was suitable for predicting the average temperature corresponding to the sample groups at hot air velocity of 0.6 m/s, and drying temperatures of 100 °C, 160 °C; hot air velocity of 1.4 m/s, and drying temperatures of 130 °C, 140 °C; hot air velocity of 2.0 m/s, and drying temperatures of 150 °C, 160 °C. The average temperature for the other sample groups was best predicted by the BP model.     

Measurement and evaluation of the apparent modulus of elasticity of apple based on Hooke’s,Hertz’s and Boussinesq’s theories

Awareness on the mechanical properties of agricultural products is necessary in order to estimate and predict the deformation of viscoelastic materials under external loads for system design, transport, processing and packaging. The aim of this study was to evaluate three theories (i.e. Hooke, Hertz and Boussinesq) on the apparent modulus of elasticity of some apple varieties i.e. Golden Delicious, Red Delicious and Granny Smith. The theoretical results were analyzed in a factorial experiment with completely randomized design with 9 treatments and 15 replicates. The results showed the practical usability of Hertz’s theory with better prediction accuracy whilst the Boussinesq’s theory showed a significant difference of predicted modulus of elasticity compared to other theories and the values reported in some publications. Although, the Hook’s theory enables the identification of a bio-yield point on its force–deformation curve, but Hertz’s theory was recommended as the most appropriate method due to easiness of working on a cylindrical sample of apple and the closer agreement with reality. Based on the results of this study, Golden Delicious and Granny Smith varieties had the lowest and the highest modulus of elasticity as 2.211 MPa and 3.431 MPa, respectively. This difference shows the firmness of apple varieties and different tissue responses to external loads and forces accordingly.     

Non-linear adsorption modeling of fatty esters and oleic estolide esters via boundary lubrication coefficient of friction measurements

The frictional behaviors of a variety of fatty esters (methyl oleate (MO), methyl palmitate (MP), methyl laurate (ML), and 2-ethylhexyl oleate (EHO)) and oleic estolide esters (methyl oleic estolide ester (ME) and 2-ethylhexyl oleic estolide ester (EHE)) as additives in hexadecane have been examined in a boundary lubrication test regime using steel contacts. Critical additive concentrations were defined and used to perform novel and simple Langmuir analyses that provide an order of adsorption energies: EHE ≥ ME > EHO > MP > MO ≥ ML. Application of Langmuir, Temkin, and Frumkin–Fowler–Guggenheim (FFG) adsorption models via non-linear fitting demonstrates the necessary inclusion of cooperative effects in the applied model. Fits of the steady-state coefficient of friction (COF)-concentration data for EHE, ME, and EHO indicate slight cooperative adsorption. MO, MP, and ML data require larger attractive interaction terms (α ≤ −2.3) to be adequately fit. Primary adsorption energies calculated via a general adsorption model are necessarily decreased while total adsorption energies correlate well with values obtained via critical concentration analyses. To account for multiple surface-site coverage a multiple-site model was defined. The intuitive assumption of multiple-site coverage of more massive components suggests deceptively increased calculated adsorption energies for typically applied models (e.g. FFG, Langmuir).     

Modified nanoporous carbon as a novel sorbent before solvent-based de-emulsification dispersive liquid–liquid microextraction for ultra-trace detection of cadmium by flame atomic absorption spectrophotometry

Combination of different extraction methods is an interesting work in the field of sample pretreatment. In the current study, for the first time, solid phase extraction combined with solvent-based de-emulsification dispersive liquid–liquid microextraction (SPE-SD-DLLME) was developed for preconcentration and trace detection of cadmium in water samples using flame atomic absorption spectrophotometry (FAAS). The adsorbed cadmium ions on prepared SPE (75 mL of aqueous solution) were eluted by optimized elution solvent and introduced to the second microextraction step. The effective variables of SPE including the pH of sample, flow rates, type, concentration and volume of the eluent and the effect of potentially interfering ions of the separation of cadmium were evaluated and optimized. Also, several factors that influence the SD-DLLME step such as pH, neocuproine concentration (the cadmium binding ligand), type of dispersed/de-emulsifier solvent, volume of disperser/de-emulsifier solvent and type and volume of extraction solvents were investigated. SPE-SD-DLLME provides a preconcentration factor of 165 for cadmium ions. Calibration plot was linear in the range of 0.1–50 μg L⁻¹ with correlation of determination (r²) of 0.988. The precision and limit of detection of proposed method were 5.1% (RSD%, n = 8) and 0.03 μg L⁻¹, respectively.     

Predicting modulus of rupture (MOR) and modulus of elasticity (MOE) of heat treated woods by artificial neural networks

In this study, MOR and MOE of the heat-treated wood were predicted by artificial neural networks (ANNs). For this purpose, samples were prepared from beech wood (Fagus orientalis Lipsky.) and spruce wood (Picea orientalis (L.) Link.). The samples were exposed to heat treatment at varying temperatures (125, 150, 175 and 200 °C) for varying durations (3, 5, 7 and 9 h). According to the results, the mean absolute percentage errors (MAPE) were determined as 0.74%, 1.01% and 1.04% in prediction of MOR values, and 1.14%, 2.21% and 2.13%, in prediction of MOE values for training, validation and testing data sets, respectively. In the prediction of MOR and MOE, values of R² were obtained greater than 0.99 for all data sets with the proposed ANN models. The results show that ANN can be used successfully for predicting MOR and MOE of heat-treated wood.     

Determination of haloperidol drug in ampoules and in urine samples using a potentiometric modified carbon paste electrode

A modified carbon paste electrode for haloperidol drug based on haloperidol-phosphomolybdate (HP-PM) as an ion-exchanger dissolved in plasticizer DBP and its potentiometric characteristics were discussed. The electrode exhibited a good Nernstian slope of 56.9 ± 0.3 mV/decade with a linear concentration range from 3.2 × 10⁻⁶ to 1.0 × 10⁻² M for the haloperidol ion. The limit of detection (LOD) was 1.5 × 10⁻⁶ M. It had response time of 5–8 seconds (s), useable in pH range of 6.2–8.6 and temperature of 20–60 °C. The electrode shows clear discrimination of haloperidol drug from several inorganic ions, sugars and some common drug excipients. The sensor was applied for determination of haloperidol drug in urine and in pharmaceutical formulations using potentiometric determination, standard addition and the calibration curve methods. The results are satisfactory with excellent percentage recovery comparable or better than those obtained by other routine methods.     

支持向量机在机床热误差建模补偿中的机理研究

针对数控机床热误差预测补偿功能,以Leaderway-V450数控机床为试验对象,通过跨季度的5批次数据,比对分析支持向量机(ε-SVR)和多元回归(MLR)两种建模方法的拟合和预测精度。研究得出,环境温度会对数控机床热误差预测模型产生干扰,模型拟合精度和预测精度在这种干扰下不具备等同性。线性ε-支持向量机在试验条件相近、环境温度接近的情况下鲁棒性较好,拟合精度、预测精度要高于多元回归算法,但当环境温度变化较大时,其预测精度较差。将跨季度环境温度条件下的数据进行综合建模,两种模型的预测精度均有所提高,但拟合精度优越的支持向量机算法建立的模型鲁棒性则低于传统的多元回归算法模型。     

Influence of crystal anisotropy on subsurface damage in ultra-precision cylindrical turning of CaF2

An optical microcavity, which stores light at a certain spot, is an essential component to realize all-optical signal processing. Single-crystal calcium fluoride (CaF₂) theoretically shows a high Q-factor which is a desirable optical property. The CaF₂ microcavity can only be manufactured by ultra-precision cylindrical turning (UPCT). The authors have studied UPCT of CaF₂ and shown the influence of crystal anisotropy and tool geometry on surface roughness and subsurface damage. The study indicated that a smaller nose radius of the cutting tool led to shallower subsurface damage. Thus, it is inferred that a smaller nose radius compared to the previous nose radius (0.05 mm) can further reduce subsurface damage. Nevertheless, the mechanism that causes a difference in subsurface damage due to crystal anisotropy is not sufficiently clear. The influence of subsurface damage on microcavity performance is still unclear. In this study, the UPCT of CaF₂ was conducted using a tool with a nose radius of 0.01 mm. The subsurface damage was investigated by transmission electron microscope (TEM) observation from the viewpoint of the change in crystal lattice arrangement. In our previous study, fast Fourier transfer (FFT) analysis was used for confirmation of change of crystal structure. In this study, FFT analysis was also used to quantitatively evaluate the depth of subsurface damage. In addition, inverse fast Fourier transfer (IFFT) was used to analyze change of crystal lattice arrangement clearly, which enables discussion of the influence of slip systems. Finally, optical microcavities are manufactured without any crack, and the influence of subsurface damage on microcavity performance is experimentally evaluated using a wavelength tunable laser and power meter.     

Ultrasonic vibration assisted grinding of hard and brittle linear micro-structured surfaces

In this paper, a novel ultrasonic vibration assisted grinding (UVAG) technique was presented for machining hard and brittle linear micro-structured surfaces. The kinematics of the UVAG for micro-structures was first analyzed by considering both the vibration trace and the topological features on the machined surface. Then, the influences of the ultrasonic vibration parameters and the tilt angle on the ground quality of micro-structured surfaces were investigated. The experimental results indicate that the introduction of ultrasonic vibration is able to improve the surface quality (The roughness SR_a was reduced to 78 nm from 136 nm), especially in guaranteeing the edge sharpness of micro-structures. By increasing the tilt angle, the surface roughness can be further reduced to 56 nm for a 59% improvement in total. By using the preferred UVAG parameters realized by orthogonal experiments, a micro cylinder array with surface roughness of less than 50 nm and edge radius of less than 1 μm was fabricated. The primary and secondary sequence of the grinding parameters obtained by the orthogonal experiments are as follows: feed rate, tilt angle of workpiece, depth of grinding, vibration frequency and amplitude. The spindle speed in the range of 1000 rpm–3000 rpm does not significantly affect the machined micro-structured surface roughness. Finally, more micro-structures including a micro V-groove array and a micro pyramid array were machined on binderless WC as well as SiC ceramic by means of the UVAG technique. The edge radius on the V-grooves and pyramids are both less than 1 μm, indicating the feasibility of UVAG in machining hard and brittle micro-structured surfaces for an improved surface quality.     

Performance analysis of a DRL-less AFE for battery-powered wearable EEG

The driven-right-leg (DRL) circuit has been commonly used in the wall-powered EEG systems to reduce common-mode interference in the bio-potential amplifier. However, DRL circuit imposes limitations on the number of channels preventing modular development, and its effectiveness is diminished for a newer generation of battery-powered EEG systems. We present a performance investigation of DRL-less EEG circuit by designing a single-channel EEG with a novel Analog Front End (AFE) that contains a differential amplifier followed by a high-Q active notch filter. The prototyped wearable EEG system has been validated to record neural signals with and without the DRL circuit. The time domain and frequency domain signals show that the designed AFE is not impacted significantly (maximum 4 dB difference) by the DRL elimination and maintains similar signal quality. The customized EEG with and without DRL offers CMRR of 72.98 dB and 71.74 dB, respectively, at 60 Hz (power-line interference range in the USA), whereas CMRR of 72.64 dB and 71.01 dB, respectively, at 20 Hz (representative EEG signal range). DRL elimination allows us to envision a sensor-level modular EEG system for neural monitoring in non-clinical environments.     

Structural and photoluminescence studies of rutile TiO2 nanorods prepared by chemical bath deposition method on Si substrates at different pH values

The effects of pH on the structural, morphological and optical properties of TiO₂ nanorods were investigated. Nanorods were fabricated on p-type (1 1 1)-oriented silicon substrates and all substrates were seeded with a TiO₂ seed layer synthesized with a radio-frequency reactive magnetron sputtering system (the TiO₂ seed layer was also examined in this research). Chemical bath deposition (CBD) was performed to grow rutile TiO₂ nanorods on Si substrate at various pH values (0.5, 0.7 and 0.9). Raman spectroscopy, X-ray diffraction (XRD) and field emission scanning electron microscopy (FESEM) analyses showed the tetragonal rutile structure of the synthesized TiO₂ nanorods. Optical properties were examined photoluminescence spectroscopy, which indicated a high-intensity UV peak centered at around 390 nm for all samples compared with visible defect peaks. Experimental results showed that the TiO₂ nanorods grown at pH 0.7 exhibited the optimal structural properties. Moreover, the CBD method enabled the formation of photosensitive and high-quality rutile TiO₂ nanorods with few defects for future optoelectronic nanodevice applications.     

Effect of pH on wear and friction of silicon nitride sliding against alumina in water

The influence of the pH in water lubricated sliding contacts was evaluated in terms of friction and wear. The experiments were carried out using a ball-on-disc setup. Si₃N₄ balls and Al₂O₃ discs were tested at temperature of (22±2) °C, sliding speed of (1.00±0.03) m/s and normal load of (54.25±0.17) N. Eight types of water with pHs varying from 3 to 12 were used as lubricant. The running-in period, friction coefficient and wear-volume were shown to be nearly independent of the initial pH values within the DLVO range (4≤pH≤10), since at these range the water׳s pH tends to the same value (7.6±0.3) during the test. Superlubricity could be reached with negligible wear by properly setting the electrochemical properties and operating conditions of the tribosystem.     

Integrated optimization using mixture design to confirm the finishing of AISI P20 using different cutting strategies and ball nose end mills

The surface roughness of a milled surface is an important response parameter in finishing milling. The quality of injected products, such as cell phones, laptops, digital cameras greatly depends on the finishing of the molds. The Ra parameter is one of the most commonly used criteria to determine the quality of milled steel. In this study, experiments were carried out on work pieces of AISI P20 steel with 20 × 20 × 10 mm. Solid carbide tools with a diameter of 6 mm were used in the experimental tests. The input parameters were the radial depth of cut, feed rate, and contact angle. The contact angle varied using a special device to simulate the several inclinations in a free-form surface. The results showed that the milled surface is a three-dimensional geometry that is influenced by many parameters. The most important input parameter for the surface roughness, however, is the feed rate, and depending on the measuring direction; the feed rate can have a significant influence on the finishing.     

Determination of particle impacts and impact energy in the erosion of X65 carbon steel using acoustic emission technique

An in-situ acoustic emission (AE) monitoring technique has been implemented in a submerged jet impingement (SIJ) system in an effort to investigate the effect of sand particle impact on the degradation mechanism of X65 carbon steel pipeline material in erosion conditions.A detailed analysis of the acoustic events' count rate enabled the number of impacts per second to be quantified for a range of flow velocities (7, 10, 15 m/s) and solid loadings (0, 50, 200, 500 mg/L) in a nitrogen-saturated solution at 50 °C. The number of impacts obtained from acoustic signals showed a strong agreement with theoretical prediction for flow velocities 7 and 10 m/s. A deviation between practical readings and theory is observed for flow velocity of 15 m/s which may be due to error from detected emissions of multiple rebounded particles.Computational fluid dynamics (CFD) was used in conjunction with particle tracking to model the impingement system and predict the velocity and impact angle distribution on the surface of the sample. Data was used to predict the kinetic energy of the impacts and was correlated with the measured AE energy and material loss from gravimetric analysis. The results demonstrate that AE is a useful technique for quantifying and predicting the erosion damage of X65 pipeline material in an erosion–corrosion environment.