Improving the SVM gender classification accuracy using clustering and incremental learning

Gender recognition has been playing a very important role in various applications such as human–computer interaction, surveillance, and security. Nonlinear support vector machines (SVMs) were investigated for the identification of gender using the Face Recognition Technology (FERET) image face database. It was shown that SVM classifiers outperform the traditional pattern classifiers (linear, quadratic, Fisher linear discriminant, and nearest neighbour). In this context, this paper aims to improve the SVM classification accuracy in the gender classification system and propose new models for a better performance. We have evaluated different SVM learning algorithms; the SVM‐radial basis function with a 5% outlier fraction outperformed other SVM classifiers. We have examined the effectiveness of different feature selection methods. AdaBoost performs better than the other feature selection methods in selecting the most discriminating features. We have proposed two classification methods that focus on training subsets of images among the training images. Method 1 combines the outcome of different classifiers based on different image subsets, whereas method 2 is based on clustering the training data and building a classifier for each cluster. Experimental results showed that both methods have increased the classification accuracy.     

Managing a hybrid energy smart grid with a renewable energy source

The aim of this work is to develop tools for optimal power flow management control in a micro grid (MG). The latter consists of a wind turbine, energy storage system, two gas turbines (GTs), and the main grid. Unlike the traditional approach, which is limited to the distribution of active power, this paper models an electrical system to coordinate and optimize the flow of both active and reactive power using discrete controls. The proposed optimal power distribution strategy has two objectives. First, it aims at forecasting over a time horizon of 24 hours the optimal distribution of the active and reactive power required for each power source connected to the MG. The proposed management incorporates the forecasts of consumption, weather, and tariffs. Second, it aims at reducing the CO₂ emissions rate by optimizing both the operating point of the two GTs and the usage of the storage unit. The proposed optimization problem for the energy management system is solved using the Bellman algorithm through dynamic programming. Its effectiveness is illustrated with various simulations carried out in the Matlab environment. The supremacy of the proposed management algorithm is highlighted by comparing its performance with conventional (restricted) management.     

Cell-Mimic Directional Cargo Transportation in a Visible-Light-Activated Colloidal Motor/Lipid Tube System

Active tether and transportation of cargoes on cytoskeletal highway enabled by molecular motors is key for accurate delivery of vesicles and organelles in the complex intracellular environment. Here, a hybrid system composed of colloidal motors and self-assembled lipid tubes is designed to mimic the subcellular traffic system in living cells. The colloidal motors, composed of gold-coated hematite, display light-activated self-propulsion tunable by the light intensity and the concentration of hydrogen peroxide fuel. Importantly, the motors show light-switchable binding with lipid cargoes and attachment to the lipid tubes, whereby the latter act as the motor highways. Upon assembly, the colloidal motor/lipid tube system demonstrates directional delivery of lipid vesicles, emulating intracellular transportation. The assembly and function of the hybrid system are rationalized by a cooperative action of light-triggered electrophoretic and hydrodynamic effects, supported by finite element analysis. A synthetic analog of the biological protein motor/cytoskeletal filament system is realized for the manipulation and delivery of different matter at the microscale, which is expected to be a promising platform for various applications in materials science, nanotechnology, microfluidics, and synthetic biology.     

An Efficient Approach for Big Data Aggregation Mechanism in Heterogeneous Wireless Connected Sensor Networks

Recently and due to the impressive growth in the amounts of transmitted data over the heterogeneous sensor networks and the emerged related technologies especially the Internet of Things in which the number of the connected devices and the data consumption are remarkably growing, big data has emerged as a widely recognized trend and is increasingly being talked about. The term big data is not only about the volume of data, but also refers to the high speed of transmission and the wide variety of information that is difficult to collect, store and process using the available classical technologies. Although the generated data by the individual sensors may not appear to be significant, all the data generated through the many sensors in the connected sensor networks are able to produce large volumes of data. Big data management imposes additional constraints on the wireless sensor networks and especially on the data aggregation process, which represents one of the essential paradigms in wireless sensor networks. Data aggregation process can represent a solution to the problem of big data by allowing data from different sources to be combined to eliminate the redundant ones and consequently reduce the amounts of data and the consumption of the available resources in the network. The main objective of this work is to propose a new approach for supporting big data in the data aggregation process in heterogeneous wireless sensor networks. The proposed approach aims to reduce the data aggregation cost in terms of energy consumption by balancing the data loads on the heterogeneous nodes. The proposal is improved by integrating the feedback control closed loop to reinforce the balance of the data aggregation load on the nodes, maintaining therefore an optimal delay and aggregation time.

     

Robust H<Subscript>2</Subscript>-PSS design based on LQG control optimized by genetic algorithms

This paper proposes a genetic algorithms (GA) optimization technique applied to power system stabilizer (PSS) for adapt a robust H₂ control based on linear quadratic controller (LQ) and Kalman Filter applied on automatic excitation control of powerful synchronous generators, to improve stability and robustness of power system type single machine connected to an infinite bus system (SMIB). Adaptation technique proposed of the robust H₂ control with the various electrical and mechanical parametric variations based on the optimization of the PSS parameters. The genetic algorithms is a search technique based on the mechanisms of natural selection of a genetic and evolution. This optimization technique is more used in the field of control for solve optimal choice problem of regulators parameters. The integration of GA to robust H₂ control with robustness test (electrical and mechanical parameters variations of the synchronous machine) show considerable improvements in dynamics performances, robustness stability and good adaptation of the robust H₂-PSS parameters under uncertain constraints. This present study was performed using our realized Graphical User Interface (GUI) developed under MATLAB.     

Integrated ion exchange mixed bed with reverse osmosis and nanofiltration for isolation of neutral dissolved organic matter from natural waters

The aim of this study was to isolate the neutral dissolved organic matter (NDOM) and the low molecular weight neutrals (LMWN) from natural waters. The coupling of an ion exchange mixed bed (IEXMB) with reverse osmosis (RO) and nanofiltration was the main hypothesis. IEXMB removed charged species, while the neutral molecules were isolated in the demineralised water and then concentrated by RO without any osmotic pressure or fouling limits. Neutrals isolation and unlimited concentration, gives this paper its originality. The nanofiltration (NF) step allows for the isolation of the LMWN. The studied reservoir water NDOM and LMWN represented respectively 35% of the dissolved organic matter (DOM) and 34% of the NDOM. Aromatic compounds were found in both fractions. The UV254 absorbance measured before and after the IEXMB evidenced the water quality ‘signature’. IEX has never been studied as fractionation method of DOM. This IEXMB approach is thus quite novel.     

Improved Geometric Anisotropic Diffusion Filter for Radiography Image Enhancement

In radiography imaging, contrast, sharpness and noise there are three fundamental factors that determine the image quality. Removing noise while preserving and sharpening image contours is a complicated task particularly for images with low contrast like radiography. This paper proposes a new anisotropic diffusion method for radiography image enhancement. The proposed method is based on the integration of geometric parameters derived from the local pixel intensity distribution in a nonlinear diffusion formulation that can concurrently perform the smoothing and the sharpening operations. The main novelty of the proposed anisotropic diffusion model is the ability to combine in one process noise reduction, edge preserving and sharpening. Experimental results using both synthetic and real welding radiography images prove the efficiency of the proposed method in comparison with other anisotropic diffusion methods.     

Experimental investigation and theoretical modeling of dehydriding process in high-pressure metal hydride hydrogen storage systems

This study explores the endothermic dehydriding (desorption) reaction that takes place in a high-pressure metal hydride (HPMH) hydrogen storage system when hydrogen gas is released to the fuel cell. The reaction is sustained by circulating warm fluid through a heat exchanger embedded in the HPMH powder. A systematic approach to modeling the dehydriding process is presented, which is validated against experimental data using two drastically different heat exchangers, one using a modular tube-fin design and the other a simpler coiled-tube design. Experiments were performed inside a 101.6-mm (4-in) diameter pressure vessel to investigate the influences of hydrogen release rate, heat exchanger fluid flow rate and fluid temperature on the dehydriding process for the HPMH Ti_1.1CrMn. It is shown the dehydriding reaction rate can be accelerated by increasing the fluid temperature and/or the rate of pressure drop. HPMH particles located in warmer locations close to heat exchanger surfaces both began and finished dehydriding earlier than particles farther away. 2-D and 3-D models were created in Fluent to assess the dehydriding performances of the modular tube-fin heat exchanger and coiled-tube heat exchanger, respectively. The models are shown to be quite accurate at predicting the spatial and temporal variations of metal hydride temperature during the dehydriding reaction.     

Consolidated method to predicting pressure drop and heat transfer coefficient for both subcooled and saturated flow boiling in micro-channel heat sinks

Published studies concerning transport phenomena in micro-channel heat sinks can be divided into those concerning saturated boiling versus those focused on subcooled boiling, with the vast majority related to the former. What has been lacking is a single generalized method to tackle both boiling regimes. The primary objective of the present paper is to construct a consolidated method to predicting transport behavior of micro-channel heat sinks incurring all possible heat transfer regimes. First, a new correlation is developed for subcooled flow boiling pressure drop that accounts for inlet subcooling, micro-channel aspect ratio, and length-to-diameter ratio. This correlation shows excellent predictive capability against subcooled HFE 7100 pressure drop data corresponding to four different micro-channel geometries. Next, a consolidated method is developed for pressure drop that is capable of tackling inlet single-phase liquid, subcooled boiling, saturated boiling, and single-phase vapor regimes as well as inlet contraction and outlet expansion. A similar consolidated method is developed to predict the heat transfer coefficient that is capable of tackling all possible combinations of heat transfer regimes. The new consolidated method is shown to be highly effective at reproducing both data and trends for HFE 7100, water and R134a.     

Coiled-tube heat exchanger for high-pressure metal hydride hydrogen storage systems – Part 2. Computational model

This second part of a two-part study presents a transient, three-dimensional numerical model for a high-pressure metal hydride (HPMH) hydrogen storage system that is cooled by a coiled-tube heat exchanger. The model uses the same geometry examined in the first part of the study and its predictions are compared to experimental results also discussed in the first part. The model involves solving coupled heat diffusion and hydriding reaction equations for Ti_1.1CrMn. These equations are solved to determine the spatial distribution of hydride temperature as a function of time over the entire duration of the hydriding reaction, which is shown to agree favorably with the experimental data. The model also serves as an effective means for tracking the detailed temporal variations of the heat exchanger’s key performance parameters for different hydride locations relative to the coolant tube. These variations can aid in determining optimum placement of the coolant tube relative the hydride powder. Like the experimental study, the model proves that coolant temperature has the greatest influence on the time needed to complete the hydriding reaction.