Showing papers in "Sadhana-academy Proceedings in Engineering Sciences in 2020"

Efficient deep learning techniques for the detection of phishing websites

Abstract: Phishing is a fraudulent practice and a form of cyber-attack designed and executed with the sole purpose of gathering sensitive information by masquerading the genuine websites Phishers fool users by replicating the original and genuine contents to reveal personal information such as security number, credit card number, password, etc There are many anti-phishing techniques such as blacklist- or whitelist-, heuristic-feature- and visual-similarity-based methods proposed as of today Modern browsers adapt to reduce the chances of users getting trapped into a vicious agenda, but still users fall as prey to phishers and end up revealing their secret information In a previous work, the authors proposed a machine learning approach based on heuristic features for phishing website detection and achieved an accuracy of 995% using 18 features In this paper, we have proposed novel phishing URL detection models using (a) Deep Neural Network (DNN), (b) Long Short-Term Memory (LSTM) and (c) Convolution Neural Network (CNN) using only 10 features of our earlier work The proposed technique achieves an accuracy of 9952% for DNN, 9957% for LSTM and 9943% for CNN The proposed techniques utilize only one third-party service feature, thus making it more robust to failure and increases the speed of phishing detection

Particle swarm optimization and feature selection for intrusion detection system

Abstract: The network traffic in the intrusion detection system (IDS) has unpredictable behaviour due to the high computational power. The complexity of the system increases; thus, it is required to investigate the enormous number of features. However, the features that are inappropriate and (or) have some noisy data severely affect the performance of the IDSs. In this study, we have performed feature selection (FS) through a random forest algorithm for reducing irrelevant attributes. It makes the underlying task of intrusion detection effective and efficient. Later, a comparative study is carried through applying different classifiers, e.g., k Nearest Neighbour (k-NN), Support Vector Machine (SVM), Logistic Regression (LR), decision tree (DT) and Naive Bayes (NB) for measuring the different IDS metrics. The particle swarm optimization (PSO) algorithm was applied on the selective features of the NSL-KDD dataset, which cut down the false alarm rate and enhanced the detection rate and the accuracy of the IDS as compared with the mentioned state-of-the-art classifiers. This study includes the accuracy, precision, false-positive rate and the detection rate as performance metrics for the IDSs. The experimental results show low computational complexity, 99.32% efficiency and 99.26% detection rate on the selected features (=10) out of a complete set (= 41).

A depth-based Indian Sign Language recognition using Microsoft Kinect

Abstract: Recognition of sign language by a system has become important to bridge the communication gap between the abled and the Hearing and Speech Impaired people. This paper introduces an efficient algorithm for translating the input hand gesture in Indian Sign Language (ISL) into meaningful English text and speech. The system captures hand gestures through Microsoft Kinect (preferred as the system performance is unaffected by the surrounding light conditions and object colour). The dataset used consists of depth and RGB images (taken using Kinect Xbox 360) with 140 unique gestures of the ISL taken from 21 subjects, which includes single-handed signs, double-handed signs and fingerspelling (signs for alphabets and numbers), totaling to 4600 images. To recognize the hand posture, the hand region is accurately segmented and hand features are extracted using Speeded Up Robust Features, Histogram of Oriented Gradients and Local Binary Patterns. The system ensembles the three feature classifiers trained using Support Vector Machine to improve the average recognition accuracy up to 71.85%. The system then translates the sequence of hand gestures recognized into the best approximate meaningful English sentences. We achieved 100% accuracy for the signs representing 9, A, F, G, H, N and P.

Numerical analysis and experimental investigation in the machining of AISI 316 steel

Abstract: High corrosion resistance and mechanical properties of AISI 316 stainless steel make its wide application in the nuclear power station and structural components in chemical industries. On the contrary, low thermal conductivity and high strain rate create problems during the machining of AISI 316, resulting in high cutting force and tool wear. Hence, this study investigates the thermal and mechanical behavior of AISI 316 steel during turning using a carbide tool. It is carried out in two stages: Finite element modeling (FEM) and experimental work. In the first stage, FEM is simulated using DEFORM software to study cutting forces, tool temperature, and chip morphology at different cutting speeds and feed rates. The results show that cutting speed and feed rate significantly affect the cutting force, thrust force and chip morphology. The chip morphology characteristics such as the degree of segmentation and serration frequency are studied. In the second stage, experimental trials are performed using the same input parameters to validate the simulated results. Results show a 10% error between simulated and experimental findings.

DVSMS: dynamic value stream mapping solution by applying IIoT

Abstract: The purpose of any business is to delight the customer as a primary stakeholder, thereby enhancing the growth and profitability. Understanding customer needs and building them on end to end value chain not only will result in serving customers on time, but also improve the effectiveness of the processes to retain competitiveness. Value stream mapping remains a popular visualization tool in the hands of the Lean Manager who seeks to produce more with less. However, value stream mapping (VSM) tends to be static and skill dependent. With the advent of Industrial Internet of Things (IIoT), there could be a paradigm shift on how VSM could be leveraged for maximizing results. IIoT makes it possible to convert the VSM as a dynamic one, enhancing with several additional parameters measured simultaneously in real time, making the relationship between cause and effect more visible. Literally, with the addition of IIoT, we could digitally re-live the moments from the past to identify the connections between the cause and effect more specifically with better accuracy. In this paper, we attempt to clarify how IIoT could enhance the VSM as a strategic differentiator for making better decisions. In a sensor-based efficiency monitoring system, the VSM becomes dynamic; thereby all the parameters including the bottleneck operations could be continuously monitored and acted upon to attain the future state eliminating the dependency on the expertise of the people.

Devanagari Handwritten Character Recognition using fine-tuned Deep Convolutional Neural Network on trivial dataset

Abstract: In order to rapidly build an automatic and precise system for image recognition and categorization, deep learning is a vital technology. Handwritten character classification also gaining more attention due to its major contribution in automation and specially to develop applications for helping visually impaired people. Here, the proposed work highlighting on fine-tuning approach and analysis of state-of-the-art Deep Convolutional Neural Network (DCNN) designed for Devanagari Handwritten characters classification. A new Devanagari handwritten characters dataset is generated which is publicly available. Datasets consist of total 5800 isolated images of 58 unique character classes: 12 vowels, 36 consonants and 10 numerals. In addition to this database, a two-stage VGG16 deep learning model is implemented to recognize those characters using two advanced adaptive gradient methods. A two-stage approach of deep learning is developed to enhance overall success of the proposed Devanagari Handwritten Character Recognition System (DHCRS). The first model achieves 94.84% testing accuracy with training loss of 0.18 on new dataset. Moreover, the second fine-tuned model requires very fewer trainable parameters and notably less training time to achieve state-of-the-art performance on a very small dataset. It achieves 96.55% testing accuracy with training loss of 0.12. We also tested the proposed model on four different benchmark datasets of isolated characters as well as digits of Indic scripts. For all the datasets tested, we achieved the promising results.

A decision support system for the selection of an additive manufacturing process using a new hybrid MCDM technique

Abstract: Recently, Additive Manufacturing (AM) has been widely used in many applications. For a particular AM component, the choice of available AM processes is critical to the component’s quality, mechanical properties, and other important factors. In that context, this article presents an efficient decision support system for the selection of an appropriate AM process. A novel hybrid Multi-Criteria Decision Making (MCDM) technique has been proposed to select an appropriate AM process from available AM processes. The Best Worst Method (BWM) is used to determine optimal weights of criteria and the Proximity Indexed Value (PIV) method is employed to rank the available AM processes. For benchmarking the abilities of an AM process, a conceptual model of spur gear was fabricated by four available AM processes viz., Vat Photopolymerization (VatPP), Material Extrusion (ME), Powder Bed Fusion (PBF), and Material Jetting (MJ). Additionally, Dimensional accuracy (A), surface roughness (R), tensile strength (S), percentage elongation (%E), heat deflection temperature (HDT), process cost (PC) and build time (BT) has been considered as most significant criteria. Further, sensitivity analysis has been performed to validate the reliability of the results. The results suggested that the Material Jetting (MJ) process produces dimensionally accurate and quality parts among available alternatives AM processes. The ranking obtained using the PIV method is consistent and reliable.

Powder-mixed electro-discharge machining performance of Inconel 718: effect of concentration of multi-walled carbon nanotube added to the dielectric media

Abstract: The present work reports an experimental investigation on Powder-Mixed Electro-Discharge Machining (PMEDM) of Inconel 718 superalloy using Multi-Walled Carbon Nanotubes (MWCNTs) dispersed in kerosene, as dielectric media. Effects of variation of peak discharge current along with concentration of carbon nanotubes in the dielectric fluid are studied in purview of machining performance indicators including material removal efficiency, tool wear rate, and surface integrity of the machined part. The obtained results are compared to that of conventional EDM which utilizes kerosene as dielectric media. Morphology and topography, these two aspects of machined surface integrity are deliberated. The following surface morphological features: uneven fusion structure, globules of debris, molten metal deposition, surface cracks, pockmarks, and recast layer are identified. Topographical study includes surface roughness, severity of surface cracking, recast layer thickness, transfer of foreign elements, surface metallurgical characteristics, residual stress, and micro-indentation hardness. It is observed that application of MWCNT mixed dielectric media substantially improves EDM performance of Inconel 718 over conventional EDM. This is due to excellent thermo-physical properties of carbon nanotubes.

Influence of textile properties on dynamic mechanical behavior of epoxy composite reinforced with woven sisal fabrics

Abstract: Due to low cost and environmentally friendly characteristics, natural fibers gain much attention over synthetic fiber. The aim of the present work is to characterize the textile properties of three different types of sisal fabric and study dynamic mechanical properties and water absorption behavior of the sisal fabric reinforced epoxy composite. Influence of grams per square meter of fabric, weaving pattern of the fabric on textile properties of the fabric is studied first. Further, the effect of the same on the dynamic mechanical properties of the sisal composites is studied. Effect of fiber weight percentage and dynamic frequency on dynamic mechanical properties also studied. Results reveal that the storage modulus (G′) decreases with increasing temperature in all the woven types of composites under consideration. However, Plain 2 (P2) and Weft Rib (WR) composites have shown better values of G′ even after the glass transition temperature (Tg). From the results, it is also evident that storage and loss modulus (G′′) increases when the yarn diameter decreases which is observed at a higher temperature also. It is also observed that fabric density also plays a significant role in the enhancement of G′ and G′′ values. The water absorption of Plain 1 (P1) based composites are found to be less compared to the other types of composites analyzed.

Optimal feature subset selection using hybrid binary Jaya optimization algorithm for text classification

Abstract: Feature selection is an important task in the high-dimensional problem of text classification. Nowadays most of the feature selection methods use the significance of optimization algorithm to select an optimal subset of feature from the high-dimensional feature space. Optimal feature subset reduces the computation cost and increases the text classifier accuracy. In this paper, we have proposed a new hybrid feature selection method based on normalized difference measure and binary Jaya optimization algorithm (NDM-BJO) to obtain the appropriate subset of optimal features from the text corpus. We have used the error rate as a minimizing objective function to measure the fitness of a solution. The nominated optimal feature subsets are evaluated using Naive Bayes and Support Vector Machine classifier with various popular benchmark text corpus datasets. The observed results have confirmed that the proposed work NDM-BJO shows auspicious improvements compared with existing work.

A novel control strategy based on hybrid instantaneous theory decoupled approach for PQ improvement in PV systems with energy storage devices and cascaded multi-level inverter

Abstract: This paper suggests an innovative control architecture based on hybrid instantaneous theory (HIT) decoupled method for improved power quality (PQ) in a photovoltaic (PV) based microgrid utilizing energy storage devices (ESD). Further, to enhance the PV-ESD performance, an eleven-level cascaded inverter (ECI) with compact size, cost, and increase in voltage level is proposed. By considering the simplicity in design and wider application, an improved perturb and observe (IP&O) method is implemented to operate the PV-ESD system at its optimum power point (OPP). In addition to that, for achieving an improved energy management, a battery-based ESD is integrated into the system. Furthermore, the use of grid LCL filter in PV is investigated with the proposed control law design to reduce the harmonic content. To verify the robustness of the HIT approach based on the harmonics and nonlinearity, various test conditions have been examined under different cases ranging from varying environmental conditions, varying grid demand and ESD charging and discharging situations by using MATLAB/Simulink software.

Probabilistic planning for participation of virtual power plants in the presence of the thermal power plants in energy and reserve markets

Abstract: Renewable energy-based on virtual power plants (VPPs) has recently attracted considerable attention for participating in energy and reserve markets due to the disadvantages of thermal power plants (TPPs). The present paper aims to maximize the VPP profitability in distribution networks including thermal power plants, at minimum load cost, using a mathematical model for implementing the VPP and evaluating its role in the energy and reserve markets. The proposed model includes a series of probabilistic scenarios used to consider the uncertainty of wind/solar generation. Therefore in the first step, the lower bound of the problem, i.e., minimizing demand cost for all the units, should be calculated. It determines the status of VPP units based on the best-case scenarios. Afterward, the problem is cut to calculate the upper bound of the problem which is maximizing the profit of the VPP. The problem is evaluated in two cases: one is the presence of VPP only in the energy market and the other is the simultaneous presence of the VPP in the reserve and energy markets. The computation ends with the convergence of lower and upper bounds of the problem. Since the proposed method uses a piece-wise model of thermal units and the problem has nonlinear equations, Mixed Integer Programming (MIP) used to calculate the contribution of units by utilizing GAMS software. Finally, the VPP profitability calculated for the day-ahead energy and reserve market after determining the method for the participation of power plants in supply at the minimum cost. The proposed method was then applied to a sample system consisting of three thermal plants, three wind farms, two solar farms, and two energy storage systems, considering several situations to examine the impact of the resources and also the resulting profitability in the energy and reserve market. The final step was the analysis of the results.

E-learning based on the adaptive learning model: case study in Serbia

Abstract: Today’s education faces many challenges related to learning and teaching efficiency, effectiveness, and costs. Contemporary research shows that the learning environment with the ability to adapt to individual needs, requirements, and competencies of students, facilitates the learning process and leads to improved learning outcomes and achievements. Nevertheless, learning management systems (LMS) that are often used in e-learning typically provide a limited level of adaptivity. The goal of this paper is to introduce an adaptive e-learning model which enables personalized learning experience and more intelligent decision making. It consists of the students’ model, the adaptation module, the expert system for data analysis and decision making, the repository of learning objects, and database of educational methods. The designed model provides adaptivity through a learning management system, considering individual characteristics of the student, such as their learning styles and prior knowledge. It is capable to adapt course content, structure, and assessment based on the specific student’s needs and performance. The model is implemented within the widely used open-source LMS, which makes it more usable and easier to deploy. The process of applying the proposed model is illustrated with a higher education case study that shows how the recommended method is applied for a successful transition to an adaptive form of learning. The model has been tested through experiment during which a group of students attended traditional non-adaptive e-learning course, and the other group attended the adaptive e-learning course. The results of data analysis showed that students who learned from an adaptive course achieved better performance in various aspects. The proposed adaptive model can enhance educational processes in terms of improving learning performance, personalized application of teaching/learning methods, as well as continuous improvement cycle.

Intrusion detection system using an optimized kernel extreme learning machine and efficient features

Abstract: In the study of Intrusion Detection System (IDS) choosing proper combination of features is of great importance. Many researchers seek to obtain appropriate features with optimization algorithms. There are several optimization algorithms that can properly select a near-optimal combination of features to reach an improved IDS. Genetic Algorithms (GA) as one of the most powerful methods have been used in this research for feature selection. In this paper, voted outputs of built models on the GA suggested features of a more recent version of KDD CUP 99 dataset, NSL KDD, based on five different labels, have been gathered as a new dataset. Kernel Extreme Learning Machine (KELM), whose parameters have been optimally set by GA, is executed on the obtained dataset and results are collected. Based on IDS criteria, our proposed method can easily outperform general classification algorithms which use all the features of the employed dataset, especially in R2L and U2R with the accuracy of 98.73% and 98.22% respectively which is the highest among the current literature.

Graph based event measurement for analyzing distributed anomalies in sensor networks

Abstract: Wireless Sensor Network (WSN) has emerged drastically with numerous practical applications of considerable Engineering importance where privacy and security are of dominant influence. This paves the way for this investigation and present interest in the development of novel and innovative intrusion detection approach. This work anticipated a novel Intrusion detection framework by modeling sensor connectivity with a targeted graph and uses statistical graph properties by modeling intrusion detection. In anticipated graph-based detection, data capturing magnitude is modeled with the Gaussian model, and the corresponding correntropy is estimated by graph matrix with adaptive sensor measurements. Anticipated detection approach is modeled based on the Laplacian Matrix, and closed-form expressions are attained for statistical analysis. At last, temporal network analysis are characterized by evaluating sensor distance among measurement distributions in consecutive time. The results depict that the anticipated detection framework offers superior detection recital than compared to existing frameworks.

Detection of electricity theft using data processing and LSTM method in distribution systems

Abstract: Electricity theft is a big problem faced by all energy distribution services and continues to rising. Therefore, studies on electricity theft detection techniques have increased in recent years. Unsuitable calibration and illegal calibration of energy meters during production may cause non-technical losses. Non-technical losses have been a major concern for the resulting security risks and the immeasurable loss of income. In most of the meter tampered locations, damaged meter terminals and/or illegal applications cannot be distinguishable during checking. In fact, electric distribution companies will never be able to eliminate electricity theft. But it is possible to take measure to detect, prevent and reduce it. In this paper, we developed by using deep learning methods on real daily electricity consumption data (Electricity consumption dataset of State Grid Corporation of China). Data reduction has been made by developing a new method to make the dataset more usable and to extract meaningful results. A Long Short-Term Memory (LSTM) based deep learning method has been developed for the dataset to be able to recognize the actual daily electricity consumption data of 2016. In order to evaluate the performance of the proposed method, the accuracy, prediction and recall metric was used by considering the five cross-fold technique. Performance of the proposed methods were found to be better than previously reported results.

Mechanical, thermal, and water absorption behaviour of jute/carbon reinforced hybrid composites

Abstract: The demand for fiber-reinforced composite materials is increasing in structural applications due to their crucial characteristics such as stiffness, strength, and durability and processing benefits at low cost. In this study, jute/carbon hybrid composite laminates were investigated for the effect of fabric hybridization and stacking sequence on tensile, impact, microhardness, water absorption, and thermal behavior of the material. The hand layup process was used to fabricate the composite laminates with four different stacking sequences. The X-ray diffraction (XRD), Fourier-Transform Infrared spectroscopy (FT-IR), Thermogravimetric analysis (TGA), and Scanning Electron Microscope (SEM) were used to characterize the structural morphology and thermal stability of the fabricated composites. The experimental results exposed that the hybridization process enhanced the properties of jute reinforced composites. FT-IR and XRD analysis revealed that the alkalization process removed the binding constituents like lignin and hemicelluloses from raw jute fiber, which resulted in a higher crystallinity index. The TGA analysis proved that the hybrid composites are thermally stable at a higher temperature. The hybrid composite with Jute/Carbon/Carbon/Jute stacking patterns has the highest tensile strength of 234.68 MPa compred to other stacking sequences. The hybrid composite with Carbon/Jute/Jute/Carbon fabric stacking sequence exhibited enhanced impact strength of 108.45 kJ/m2 and better moisture resistance. The incorporation of jute with carbon declined the tensile strength and impact strength by 22% and 14%, respectively, compared to carbon-reinforced composites. The surface micrographs of the fractured samples exhibit the interfacial bonding of fiber/matrix, matrix crack, fiber fracture, and fiber pullouts.

Electromyogram (EMG) based fingers movement recognition using sparse filtering of wavelet packet coefficients

Abstract: Surface electromyogram (EMG) signals collected from amputee’s residual limb have been utilized to control the prosthetic limb movements for many years. The extensive research has been carried out to classify arm and hand movements by many researchers. However, for control of the more dexterous prosthetic hand, controlling of single and multiple prosthetic fingers needs to be focused. The classification of single and multiple finger movements is challenging as the large number of EMG electrodes/channels are required to classify more number of movement classes. Also the misclassification rate increases significantly with the increased number of finger movements. To enable such control, the most informative and discriminative feature set which can accurately differentiate between different finger movements must be extracted. This work proposes an accurate and novel scheme for feature set extraction and projection based on Sparse Filtering of wavelet packet coefficients. Unlike the existing feature extraction-projection techniques, the proposed method can classify a large number of single and multiple finger movements accurately with reduced hardware complexity. The proposed method is compared to other combinations of feature extraction-reduction methods and validated on EMG dataset collected from eight subjects performing 15 different finger movements. The experimental results show the importance of the proposed scheme in comparison with existing feature extraction-projection schemes with an average accuracy of 99.52% ± 0.6%. The results also indicate that the subset of five EMG channels delivers similar accuracy (>99%) to those obtained from all eight channels. The resultant accuracy values are improved over the existing one reported in the literature, whereas only one-third numbers of channels per identified motions are employed. The experimental results and analysis of variance tests (p < 0.001) prove the feasibility of the proposed work.

Modelling and multiresponse optimization for minimizing burr height, thrust force and surface roughness in drilling of ferritic stainless steel

Abstract: Although there have been many studies on the drillability of various grades of stainless steel, there is no scientific research on the drilling of ferritic stainless steel. Also, the burr at hole exit means the need for secondary machining operation and indirectly increases the production cost. Thus, this study focused on the modeling and minimizing burr height (Bh), thrust force (Fz) and surface roughness (Ra) during drilling of AISI 430 ferritic stainless steel with uncoated carbide drill under dry condition. Bh, Fz and Ra based on different cutting speed and feed rates were measured during drilling tests, and then cutting parameters were optimized by applying Taguchi based grey relational analysis. Moreover, the mathematical models were created by employing the response surface method to predict the machining outputs. The thrust force and the surface roughness decreased while the burr height increased with the increase in cutting speed. Uniform burr formation with drill cap was observed for all machining parameters under dry environment. The effect levels of feed rate and cutting speed on burr height were determined as 54.82% and 44.67%, respectively. These result shows that cutting speed is as important as the feed rate during the drilling of the ferritic stainless steel. In the current study, the best suitable levels of feed rate and cutting speed were detected as 0.12 mm/rev and 45 m/min for minimizing Bh, Fz and Ra. The coefficients of determination obtained by RSM indicated a relationship in high level between the cutting parameters and machining outputs.

Single and multiple odor source localization using hybrid nature-inspired algorithm

Abstract: In this paper, optimization-based approach has been adopted to localize the odor source in an unknown environment. Two scenarios taken into consideration, first single odor source (SOS) with a point source emission at a constant rate and four multiple odor sources (MOS) with point source emissions and different release rates constant in time. In context to SOS, four environments that have distinct dimensional layout have been generated with slight variation in wind velocity and diffusion constant. In case of MOS, there are five environments with same layout but different contributing factors such as wind velocity, placement of odor sources and emission rates which are considered to demonstrate its impact on success rate of algorithms. A recent optimization technique called hybrid teaching learning particle swarm optimization (HTLPSO) has been adopted and implemented in all the arenas, namely SOS and MOS, where mobile robots AKA virtual agents (VAs) are working in collaboration. There are group of VAs deployed in this operation ranging from {3–15}. To investigate the effectiveness of the algorithm, results of HTLPSO are compared with classical particle swarm optimization (PSO) and teaching learning-based optimization (TLBO). It is observed that HTLPSO outperforms TLBO and PSO in arenas with larger dimensions while utilizing few iterations in comparison with other algorithms in case of SOS. HTLPSO also performs best in case of MOS, surviving the effect of wind velocity and change in emission rates. Only when odor sources are placed differently and scattered, TLBO gives the best result. Another highlight of HTLPSO is convergence with high accuracy even with less number of VAs.

On the static deformation of FG sandwich beams curved in elevation using a new higher order beam theory

Abstract: This article presents the static analysis of FG sandwich beams curved in elevation. Navier-type semi-analytical solutions are obtained based on polynomial type fifth order shear and normal deformation theory. The beam has FG skins and isotropic core. Material properties of FG skins are graded in z-direction according to the power-law distribution. The present theory accounts for a fifth-order distribution of axial displacement and fourth-order distribution of transverse displacement. The present theory considers the effect of thickness stretching and gives a realistic variation of transverse shear stress through the thickness of the beam. The governing equations are obtained within the framework of the principle of virtual work. Semi-analytical static solutions for the simply supported FG sandwich beams curved in elevation are obtained using Navier’s technique. The beam is subjected to uniformly distributed load. The non-dimensional numerical values for displacements and stresses are obtained for various power-law index and thickness of the core. The present results are found in good agreement with previously published results.

Experimental investigation on the filtration characteristics of a commercial diesel filter operated with raw and processed karanja-diesel blends

Abstract: With soaring fuel prices, engine emissions, and demand for energy security, the development of alternative fuels from renewable resources has become significant. The current work explores the filtration performance of a commercial diesel filter operated with diesel blends of raw and processed karanja oils. Three processes viz. esterification, transesterification, and degumming were performed with karanja oil, and their diesel blends (20% v/v) were tested in a continuous run filtration system. The fuel filter plays a significant role in removing the contaminants/impurities from the fuel. Hence, the pressure difference across the fuel filter was monitored at the flow rates (0.03 m3/h and 0.075 m3/h) and fuel temperatures (30°C and 70°C). Besides, the effects of weight gain and surface morphology on the filter characteristics were observed during the study. Further, the head loss factors of fuel filter under varying operating conditions were deduced from the experimental measurements. The results of different processed oil blends were compared with fossil diesel to find the suitable blend. By considering the fossil diesel as the reference case, the average pressure difference for the degummed blends was significantly lower compared to other tested blends, which highlights the importance of the degumming process. Furthermore, the tested images of filter revealed the presence of minor components such as sludge or gums on the filter surface for the fuel without the degumming process.

Investigations on microstructural and microhardness developments in sintered iron–coal fly ash composites

Abstract: The present work is aimed to explore the microstructural and mechanical characteristics of coal-fly ash reinforced iron metal-matrix composites (IMMCs), synthesized through powder metallurgy technique. Coal-fly ash wt%, compacting load and sintering temperature were considered as the input variables, whereas sintered density and microhardness of the composites were taken as the output responses. Flowability and compressibility of the starting materials were demonstrated using Hausner ratio and Carr’s index. Decorous morphological, crystallographic and elemental characteristics of the starting materials and IMMCs were deliberated using Scanning electron microscopy, X-ray diffraction and Energy-dispersive X-ray spectroscopy investigations respectively. A significant improvement in the microhardness of IMMCs by 50% and drop in density by 35% were found at 15 wt% as compared to 0 wt% reinforcement. The substantial increase in the microhardness eventually resulted in an increase in their specific microhardness by a factor of two. Significant improvements in the microhardness of IMMCs at 15 wt % of reinforcement, compacted at 10 ton and sintered at 1150°C were found to be prompted by the strengthening mechanisms like load transfer, Hall–Petch effect and Taylor strengthening. The analytically calculated microhardness in the light of strengthening mechanisms was found smaller than the corresponding experimental values as a function of wt % of reinforcement. Further, statistical analysis of the obtained results was carried out using response surface methodology.

Design, analysis, fabrication, control and comparative study of two different-shaped plate levitation prototypes

Abstract: This paper presents a comparative study on the design, modelling, electromagnetic analysis based on finite-element software, fabrication and experiment on rectangular flat (148 g) and C-shaped (148 g) levitation prototypes based on steel plates. No mechanical restrainer has been used in the transverse direction for the levitation. This aspect of the work is an improvement over existing work reported in the published literature. The entire set-up has been designed, fabricated, analytically investigated and experimentally evaluated and verified. The finite-element model (FEM) has been derived using standard commercial package(s). The analytical model has been obtained using specific permeance concepts following Robert Pohl’s method. Excellent correlation between the predicted and experimental results is a highlight of the work. The stability against transverse mechanical perturbation has also been investigated. Control system design and implementation is successfully done.

SDN-based DDoS Attack Mitigation Scheme using Convolution Recursively Enhanced Self Organizing Maps

Abstract: In a cloud computing environment, the Distributed Denial of Service (DDoS) attack is considered as the crucial issue that needs to be addressed in ensuring the availability of resources that emerge due to the compromisation of hosts. The process of detecting and preventing DDoS attacks is determined to be predominant when the potential benefits of decoupling data plane from the control plane are facilitated through the Software Defined Networking (SDN) in the cloud environment. The incorporation of SDN in DDoS mitigation also enhances the probability of investigating the data traffic flow using the reactive process of updating forwarding rules, analyzing the network with a global view and centralized control in monitoring for better DDoS mitigation enforcement. In this paper, a Convolution Recursively Enhanced Self Organizing Map and Software Defined Networking-based Mitigation Scheme (CRESOM-SDNMS) is proposed for ensuring the better rate of detection during the process of preventing DDoS attacks in clouds. This proposed CRESOM-SDNMS facilitates a predominant option in resolving the issue of vector quantization with enhanced topology preservation and the superior initialization mechanism during the process of SOM-based categorization of flooded data traffic flows into genuine and malicious. The simulation experiments and results of the proposed CRESOM-SDNMS confirmed a superior classification accuracy of around 21% when compared to the existing systems with minimized False Positive rate of 19% compared to the benchmarked DDoS mitigation schemes of the literature.

Heterogeneous classifier ensemble for sentiment analysis of Bengali and Hindi tweets

Abstract: Sentiment analysis is an essential step for analysing social media texts such as tweets and other posts on the various micro-blogging sites. The basic step of sentiment analysis is sentiment polarity detection, which identifies whether an input piece of social media text is positive, negative or neutral. In this paper, we present an approach that combines heterogeneous classifiers in an ensemble for sentiment polarity detection in Bengali and Hindi tweets. Our proposed method constructs an ensemble of three different base classifiers where the feature set for each base classifier is different from each other. We have also incorporated an external knowledge base called sentiment lexicon to augment tweet words with sentiment polarity information retrieved from the sentiment lexicon. Experimental results show the effectiveness of our proposed heterogeneous ensemble model for sentiment polarity detection for both Bengali and Hindi languages. It has been shown that our system outperforms other existing Bengali and Hindi sentiment classification systems to which it is compared.

Performance analysis of current lightweight stream ciphers for constrained environments

Abstract: Maintaining an adequate balance between security and other performance metrics like memory requirement, throughput and energy requirement in a resource-constrained environment is a major challenge. The National Institute of Standards and Technology (NIST), in its latest lightweight cryptography report, suggested the suitability of symmetric ciphers in constrained devices. In this paper we have performed statistical security analyses of six state-of-the-art stream ciphers, namely Lizard, Fruit, Plantlet, Sprout, Grain v1 and Espresso, with the help of randomness test, structural test, autocorrelation test and avalanche test. We have also carried out the performance analysis of these ciphers in detail after porting the optimized code of the ciphers to a low-cost microcontroller, namely ATmega 328P. The selection of the device is based on its acceptability in the Internet of Things (IoT)-based network. The statistical security, performance metrics and comparative analysis suggest the suitability of the selected ciphers for providing security in constrained environments.

Detection of situational information from Twitter during disaster using deep learning models

Abstract: Twitter is an excellent resource for communicating between the victims and organizations during a disaster. People share opinions, sympathies, situational information, etc., in the form of tweets during a disaster. Detecting the situational tweets is a challenging task, which is very helpful to both humanitarian organizations and victims. There is a chance that both situational and non-situational information may be present in a tweet. Most of the existing works focused on identifying single-information-type tweets like situational information, actionable information, useful information, etc. Detecting the mixture of situational and non-situational information tweets remains a challenging task. Although existing works designed an SVM classifier using low-level lexical and syntactic features for classifying situational and non-situational tweets, their method does not work well for a mixture of situational and non-situational information tweets. This paper addresses the problem of detecting the situational tweets using different deep learning architectures such as Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), Bi-directional Long Short-Term Memory (BLSTM) and Bi-directional Long Short-Term Memory with attention (BLSTM attention). Moreover, deep learning models are applied to Hindi language tweets besides English language tweets for identifying the situational information during a disaster. Some of the tweets are posted in the Hindi language, where the information is not available in the English language in countries like India during the disaster. Experiments are performed on various disaster datasets such as Hagupit cyclone, Hyderabad bomb blast, Sandhy shooting, Nepal Earthquake and Harda rail accident in both in-domain and cross-domain. The results of deep learning models demonstrate that it outperforms the existing traditional approach, such as the SVM classifier with low-level lexical and syntactic features for detecting the situational tweets during the disaster. Additionally, to our best knowledge, this is the first attempt in applying the deep learning models to identify the Hindi language situational tweets during the disaster.

Photodegradation of Methylene Blue (MB) using Cerium-doped Zinc Oxide nanoparticles

Abstract: Cerium-doped Zinc Oxide nanoparticles were prepared with co-precipitation method. X-Ray diffraction (XRD) analysis shows sharp peaks conforming the crystalline nature of prepared particles. Crystallite size was found from XRD peaks. A Field Emission Scanning Electron Microscope (FESEM) was used to see the particles and hexagonal shape. Spectroscopic properties were studied using a Fourier Transform-Infra Red (FT-IR) spectrometer and Raman spectrometer. Optical properties were studied using a Diffuse Reflectance Ultraviolet Visible (DRS UV–VIS), spectrophotometer. Band gap of doped particles reduces as doping concentration increases. Reduced band gap of Cerium doped particles makes them a good catalyst. There are many organic dyes which are used as coloring agents by industries like textile, printing, etc. These industries release harm full pollutants into water bodies, which effect human and aquatic life. Prepared Cerium-doped ZnO nanoparticles have proved to photodegrade Methylene Blue. Undoped ZnO nanoparticles decolorized the solution in 210 min with 81.93% efficiency. Ce (0.03 wt%) doped ZnO nanoparticles decolorized the solution in 120 min with 92.62% efficiency.

Performance analysis of a vortex chamber under non-reacting and reacting conditions

Abstract: A series of non-reacting and reacting flow experiments are performed in a vortex combustion cold-wall (VCCW) chamber using gaseous oxygen and gaseous hydrogen as propellants. Oxidizer is injected tangentially at the aft end of a combustion chamber from four ports. Hydrogen is injected axially from the top-centre of the chamber. The oxidizer to fuel mixture ratios considered for the experimental studies are in the range of 4.2–6.0 for non-reacting case, and 6.38 for reacting flow experiments. Numerical simulations under non-reacting conditions are conducted to understand the flow behaviour in the chamber at a mixture ratio of 4.2 considering the same propellants used in the experiment. Results from non-reacting flow cases indicated that the chamber pressure increased by 0.8 bar with an increase in the mixture ratio from 4.2 to 6.0. The chamber pressure developed under the reacting flow conditions is found to be higher by around 1.3 bar compared with the non-reacting flow condition. The oxidizer concentration is found to be higher along the inner chamber wall, thus limiting the wall surface temperature to 360 K in the reacting conditions.