Search

This research aimed to study the optimal condition of natural color extraction from the bracts of Bougainvillea glabra and to develop a cosmetic product containing color lake from the red-purple bracts. The dye extraction was optimized for three conditions (solvent concentration, solid-liquid ratio, and extraction time) and analyzed by spectroscopy. The obtained dye was synthesized into lake pigments in 7 different salts. Then, the red-purple lake was investigated by colorimeter and Fourier transform infrared spectroscopy (FTIR). Lipstick containing the pigment was developed. The optimal condition for dye extraction from the bracts was with 50% ethanol, a solid-liquid ratio of 1:30 g/mL, and an extraction time of 30 minutes. Interestingly, the paper flower lake pigment was only prepared from aluminum chloride. The whiteness (L), redness (a), and yellowness (b) of the aluminum lake were 37.51 (±1.62), 10.23 (±0.12), and 12.65 (±0.22), respectively. The infrared spectrum of the aluminum lake sample appeared at a wavenumber of 1624.38 cm-1 that indicated carbonyl group (C=O stretching) connected with amide bonds. All of the lipstick formulations containing aluminum lake in this experiment were not sweaty. The lipstick consisting of 1% red-purple lake had the closest color with the master formula (∆E=1.24), aand its melting point was 65.3 (±0.3) °C. In conclusion, this research found a novel cosmetic ingredient to overcome the lipstick sweating problem.

This paper reports the development of hybrid monitoring and analysis software for those who are interested in hybrid analysis. When the machines were operating, the situations that made motor current, vibration and oil could demonstrate the characteristics of machinery status. Somehow, to provide engineers could detect and make maintenance plans by observing and collecting the data as speed and magnitude to compare with ISO standards. This may be of benefit in many industries such as aerospace and industrial power plant etc. In the experiment, the software may collect the data from a sensor through NI hardware. To experiment the project for testing by using the virtual machine, the experiment has gone through the motor rotation and add the mass on shaft to compare the data and distinguish the status that shows upon interface the program. In this paper, the performance was analysed based on the RMSE between fuzzy logic and Neuro-fuzzy systems on MATLAB. The result errors obtained from fuzzy logic of 70% to 140% were relatively higher if compared to Neuro-fuzzy systems of 3% to 37% due to Neuro-fuzzy training from real hybrid data but fuzzy systems designed reference follow ISO standard.

Human activity recognition using streaming data from the accelerometer sensor of smartphone is still an interesting issue for researchers. Most researches develop the recognition model based on personal model type which require the training data obtained from only user who will utilize the model. To prepare the training data, the user must perform various activities and annotate them within the specified time. This is a major inconvenience for the users. In this paper, we propose a new smartphone-based dynamic framework for physical activity recognition named “ISAR+”. The new framework is an impersonal (universal) model which can be built once and used on new users without requiring labeled training data from those users. Because the proposed model is adaptability with evolving data streams of each new user by using the incremental learning for real-time recognition. This work was validated the proposed model in terms of prediction accuracy and usage times on two public activity recognition datasets. The experimental results

Milk collection centers collect raw milk from farmers and measure both its quantity and quality. Both data are used to calculate payments to farmers. Daily operations at the center include quality assessment, i.e. alcohol test, methylene blue test and somatic cell count. Laboratory assessment includes pasteurized count, total solid and somatic cell count. At present, the results of assessment are manually recorded on paper and are then inserted into an Excel file for money calculation. The serious problem with this method is that their errors occur during manual recording, delaying the calculation process because the workers must trace back and correct the errors. This research aims at 1) developing an innovation for milk collection centers by deploying information technology, and 2) reduce error and time during the data collection. We developed an information system following Model View Control (MVC) architecture, which allows for rapid development of the system, and, more importantly, support further scalability in the future. The developed system provides both convenience and precision in the data recording process, and helps reduce errors and time consumption significantly.

The Satin Bowerbird Optimization (SBO) is a recently developed meta-heuristic optimization algorithm which was inspired by the Satin Bowerbirds living in Australia’s rainforests and other mesic habitats. Like other metaheuristic algorithms, the main problem faced by the SBO is that it has been empirically demonstrated to become easily trapped into local optimal solutions, creating low precision and slow convergence speeds. Therefore, in an effort to enhance global convergence speeds and in order to obtain a better performance, this paper introduces the chaos theory into the SBO optimization process. Various chaotic maps were considered in the proposed Chaotic-SBO (CSBO) method in order to replace the main parameter’s greatest step size (), which assists in controlling both exploration and exploitation. The proposed CSBO methods are benchmarked within the CEC2014 test problems. The numerical results show that these novel algorithms improve the chaotic maps, especially in the Tent map, as well as improving the performance of the original Satin Bowerbird Optimization Algorithm. Specifically, the CSBO10(Tent) achieved the best performance.

The aim of this paper is to study properties of connectedness in ideal generalized topological spaces. Precisely, we provide characterization of -connected sets, connected sets, -separated sets and -component.

This paper compares the performance of speed responsesgiven by the direct vector control system of an induction motor. The four PI controllers are incorporated in such a control system with eight state-variables. The coefficient matrix of the above closed-loop system holds eight eigenvalues. The direct computation method and random selection was used to specify all the PI gains. The direct method proceeds with a decision on eight appropriate negative integers of eigenvalues and then solving a set of eight simultaneous polynomial-equations. All its roots become proper PI gains. The random selection assigns arbitrarily each PI gain to a positive integer.When all the parameters of the induction motor and its mechanical load match with the corresponding nominal values and when some parameters deviate from the corresponding ones, the direct method providesspeed responses preferable to those caused by random selection. Five instances are included in simulation. These involve matching between each parameter and the corresponding nominal value, atenfold increase of either rotor resistance, stator resistance, or moment of inertia from the corresponding one, andsimultaneously a fourfold increase of rotor resistance, stator resistance, and moment of inertia from the corresponding ones.

This paper demonstrates design of a commercial aircraft at the conceptual design stage by considering 2 objective functions i.e. minimization of aircraft maximum takeoff weight and maximization of lift-to-drag ratio multi objective evolutionary algorithms (MOEAs). Dynamic stability is taken as a design constraint while design variables include geometrical properties of the plane, consisting of the wing, empennage and fuselage. The aerodynamic coefficients of flight were calculated by using SUAVE (Aerospace Conceptual Design Environment) whereas dynamic stability analysis was achieved through using Athena vortex lattice (AVL) software. Pareto front obtained from various MOEAs are compared by Hypervolume. MOEAs are contained as MOEA / D, Two-Arch2, RPBILDE, PICEA-g, KnEA/A-ENS and KnEA. The obtained solutions for aircraft can be used for further design in the preliminary design stage.

This paper aims to present a calculation of energy levels of (12-6) Lennard-Jones potential of the quantum mechanical system of the bound state problem by using finite difference methods (FDMs) with a truncation error for constructing a proper Hamiltonian matrix and calculating the eigenvalues and eigenvectors from this matrix. The interaction potential between two-particle system in the atom is in the kind of attraction and repulsion. As a result, we represent the energy in each level through a unitless energy parameter and the radial probability distribution of energy levels is also presented. Moreover, we also illustrate graphs of radial probability distribution with respect to the distance in each energy level.

This paper examines a deep convolutional neural network (Deep CNN) for plant recognition in the natural environment. The primary objective was to compare 4 CNN architectures including LeNet-5, AlexNet, GoogLeNet, and VGGNet on three plant datasets; PNE, 102 Flower, and Folio. The images in the PNE and 102 Flower dataset include a complicated background because they were taken in a natural environment. On the other hand, the images in the Folid dataset are only leaf images that were taken in a laboratory environment using a white background. The comparison of deep CNN using GoogLeNet and VGGNet Architecture show that GoogLeNet outperformed while working on the PNE and 102 Flower dataset when using a training time with iterations of 10,000 epochs. GoogLeNet also faster than the VGGNet architecture. However, the experiment showed that the VGGNet architecture outperforms the other CNN architectures on the Folio dataset and used only 1,000 epochs for training. In our experiment, we can create a model from the deep CNN using GoogleNet architecture, and this is because it showed better results with the plant images that were taken in the natural environment.

Plagiarism is deemed a serious ethical offense. Without proper protection and accurate detection algorithms, the number of innovations would significantly dwindle, impeding the advancement of national research and education. Especially in programming courses, plagiarism is extremely common. Most students tend to do their homework in the simplest way: copying the source code of others. Such unpleasant issues must be addressed to preclude this misbehavior before it becomes students’ habit, and to promote positive attitudes in our juveniles. In this paper, we, propose an algorithm to identify original source codes, helping instructors to pinpoint whom to monitor and be able to set measures or correctly warn each wrong-doing student. This work introduces a rule, derived from a m+l evolutionary algorithm, for identifying which source codes are original. First, we run the maximum spanning tree algorithm to generate a phylogenetic tree, representing the relationships and sequences of plagiarisms.Then, we apply our proposed rule to identify the source and its copies, whose relations are represented by arrows’ directions. The experiment result shows that the proposed rule yields the accuracy of 90.24%.

Nowadays, most universities provide a learning management system (LMS) to support teaching and learning. These LMS provide an ability to publish learning content, announce news and exchange opinion etc. However, most students may not pay attention to LMS when compared with online social media which play a key role in student’s everyday life. This paper presents the development of a Facebook application called F-learning for supporting Moodle LMS by presenting activities such as quiz and assignment files. F learning will notify students through Facebook when activities have been updated. Moreover, F-learning is also able to rank quiz score by H-index concept, rewarding the top-3 in each course, displaying and sharing overall medal table in accordance with the Gamification concept. The evaluation results reveal that F-learning can motivate students to use the learning manage system.

In this paper, factors affecting mechanical properties and metallurgical structure of semi-solid aluminum alloy grades 356 and 7075 were investigated by a transient liquid-phase bonding method. ZA 27 zinc alloys were used as coagulants (solders) in a liquid state during the welding process. Semi-rigid aluminum castings with butt joints were prepared as a cylindrical shape 12 mm in diameter, 45 mm length and ZA 27 zinc alloys were 12 mm in diameter, and 0.25 mm thick. Variables in this transient liquid-phase bonding method included 60 minutes and 120 minutes welding time, welding temperatures at 450, 500 and 550 degrees Celsius, compression pressure at 4 MPa, and the flow rate of argon gas atmosphere at 4 liters/min. According to the results, the physical characteristics of all samples were well bonded. The microstructure in the bond line showed that size of the grain had grown because of the heat during transient liquid-phase bonding. However, at temperature 550ºC and compressive holding time for 60 minutes some porosities showed after welding, and the tensile strength of all samples were broken in the bonded line. However, at temperature 500ºC and compressive holding time 60 minutes, the maximum average tensile strength at 103.06 MPa was in the bonded line area where the hardness properties were higher than in other areas. In short, transient liquid-phase bonding technique could reduce welding time when compared to the diffusion bonding process.

In this paper, we study the properties of open sets, closed sets, closure operator and interior operator on minimal structure space. We will provide characterization of sg -submaximal space by using various kinds of generalized closed sets and open sets.