Showing papers presented at "IEEE India Conference in 2018"

Optimal Load-frequency Regulation of BioRenewable Cogeneration based Interconnected Hybrid Microgrids with Demand Response Support

Abstract: This work is an earliest attempt to study the load-frequency regulation of Wind, Solar-thermal, Microhydro, Biogas, and Biodiesel generating unit based interconnected hybrid microgrids with Demand response support. The linearized models of each renewable units are established for the proposed interconnected two-unequal hybrid microgrid system along with demand response strategies. The load-frequency responses of the proposed system are studied using Particle swarm optimization tuned classical PID controllers for different scenarios of source and load variations. Initially, the responses of system with microhydro and biogas units, are studied, subsequently connecting Wind and Solar-thermal units, which witness the increment in oscillations with penetration of renewable units. Then, the oscillations are reduced optimally by the inclusion of biodiesel generator and demand response supports. Finally, the responses are studied for simultaneous variations in renewable-sources and load-demands, reporting satisfactory regulation of load frequency with demand response contributions.

3D Ear Biometrics: Acquisition and Recognition

Abstract: Ear is one of the recent biometric traits used to recognize the human. The ear can be considered as a robust biometric feature with cosmetics, age, and facial expressions, which is unlikely with the other biometrics features. The ear has a predictable background and moreover, the ear can be used an effective supplementary biometric trait with various other biometrics. Furthermore, the distinctiveness property of the ear is highly motivated for security related operations. In this paper, a method for 3D ear acquisition is presented, which is used in creating one of the largest 2D and 3D ear database (IIT Indore) and the recognition performance on the acquired data using a 3D descriptor. We have used Signature of histograms of orientations (SHOT) and Binary SHOT (B-SHOT) descriptor as a 3D feature descriptor and matched the query and repository images and obtain a rank one authentication rate of 97.13% and 97.12%, respectively.

A Hybrid Evaluation Framework for QoS Based Service Selection and Ranking in Cloud Environment

Abstract: With the growing demand and the varied availability of a large number of cloud services, it has become difficult to discover a relevant cloud service that can satisfy cloud customer’s functional needs. From the customer’s point of view, it is hard to decide, which cloud service satisfy their QoS (Quality of Service) requirement and what is the basis for their selection. To assist the cloud customer in the optimal service selection, a framework has been proposed, which allows cloud customer to assess the many cloud service providers by considering the QoS attribute value. In this work, we put forward an improved MCDM (Multi Criteria Decision Making) methodology that improves the existing cloud service selection methods. In particular, we use Best-Worst method for calculating the attribute weights and apply TOPSIS (Technique for Order Preference by Similarity to Ideal Solution) method for calculating the final rank of cloud service provider. Furthermore, we use a real time case study to analyze the pertinence of the proposed technique. The experimental results affirm the feasibility and effectiveness of the proposed methodology. This research helps to create a strong competition among the cloud service providers.

Hand Gesture Recognition Using Deep Network Through Trajectory-to-Contour Based Images

Abstract: Vision-based hand gesture recognition involves visual analysis of hand shape, position and/or movement. Most of the previous approaches require complex gesture representation as well as selection of robust features for proper gesture recognition. In this paper, a model-based method for hand gesture recognition has been presented using convolutional neural network. The model is fed with trajectory-to-contour based images obtained from isolated trajectory gesture through segmentation and tracking the hand motion, thereby estimating the hand motion trajectory for recognition. Conventional methods can extract low-level features, while deep learning approaches learn image features hierarchically from local to global with multiple layers of abstraction from vast number of sample images. Feature learning capability of CNN architecture has given outstanding results on three different datasets.

Isolated Multiport Converter for Solar PV Systems and Energy Storage Systems for DC Microgrid

Abstract: Multiport converters (MPC) are the cost-effective and efficient solution for integrating multiple energy sources and loads. This paper presents an isolated four port bidirectional multiport structure for interfacing Solar PV and storage devices with DC microgrid (DCMG). Solar PV is used as the preliminary energy source whereas the battery along with the supercapacitor (SC) work as an energy buffer for the system. The converter is based on a triple active bridge, which is coupled with an additional half bridge bidirectional switching cells to form the four port topology. In the developed scheme, a decoupled phase-shifted control strategy is employed to control the power flow between the ports. The decoupled control strategy eliminates the manipulation of possible interdependency in the control of different converter ports. The MPC and the control strategy has been validated through simulation studies under various conditions.

A Sequence-to-Sequence Model Approach for ImageCLEF 2018 Medical Domain Visual Question Answering

Abstract: Numerous attempts have been made in the recent past for the task of free-form and open-ended Visual Question Answering (VQA). Solving VQA problem typically requires techniques from both computer vision for a deeper understanding of the images and Natural language processing for understanding the semantics of the question and generating appropriate answers. It has caught the attention of a lot of researchers because of its enormous applications in the real-world scenarios. But none of the existing approaches are designed for the medical image-question pairs which require a sequence of words as an answer. We propose a novel approach by combining the tasks of Image captioning and Machine translation and provided a comprehensive model that takes a medical image-question pair as an input and generates a sequence of words as an answer. We evaluate our model on the dataset provided by ImageCLEF as a part of the ImageCLEF 2018 VQA-med challenge. We outperformed all the contestants of the challenge by achieving the best BLEU and WBSS scores. Furthermore, we provide additional insights that can be adopted to develop our baseline model and the challenges that lie ahead of us while building Machine learning models for medical datasets.

Lattice Reduction Assisted Likelihood Ascent Search Algorithm for Multiuser Detection in Massive MIMO System

Abstract: Massive multiple input multiple output (MIMO) system achieves high spectral and energy efficiency by incorporating a large number of antennas at the transmitter and/or receivers. Multiuser detection is an important task that needs to be done at the receiver of the Massive MIMO system to mitigate multiuser interference. The classical Zero Forcing (ZF) detector suffers from high residual interference. By making channel matrix orthogonal, the Lattice Reduction (LR) techniques can be assisted for the ZF detector to minimize interference. On the other hand, the Likelihood Ascent Search (LAS) is a neighborhood search based low complexity detection algorithm that is used for massive MIMO systems. It takes the Zero Forcing (ZF) solution as initial vector and searches for a near-optimal solution by examining cost values of its neighborhood vectors. The performance of the LAS algorithm is mainly relying on an initial vector. So, this paper investigates the design of LAS algorithm with LR assisted ZF solution as an initial vector to improve performance over the classical ZFLAS detector. The proposed algorithm attains a more achievable trade-off between Bit Error Rate (BER) performance and exponential time detection complexity for large extended systems.

FPGA Based System Integration Scheme for a Grid Connected Photovoltaic System

Abstract: This paper presents development of a Quadrature Axis Hysteresis Current Controller (QAHC) for a three phase single stage grid connected PV (GCPV) system on a Field Programmable Gate Array (FPGA) in view of achieving simple but effective real-time implementation. The quadrature axis component of the inverter current is used to control the injected inverter current to the grid for achieving unity power factor further. Further, the inverter current is controlled to extract the maximum power from the photovoltaic (PV) modules. We present hardware implementation of measurement systems and FPGA based reconfigurable controller design in parallel programming VHSIC Hardware Description Language (VHDL) for a three phase GCPV system. Experimental results confirm the accuracy and stability of the proposed QAHC based system integration for the studied GCPV system.

Localization Using Stochastic Gradient Descent Method in a 5G Network

Abstract: In this paper, we propose an iterative method for smartphone localization in a 5G network. The location estimation accuracy degrades for an inbuilt GPS smartphone in dense environments and indoor scenarios. We propose a combined model that uses the Received Signal Strength (RSS) from macrocells and femtocells thereby increasing the localization accuracy. The location is estimated by optimizing the path loss model for Macro Base Station (MBS) and femtocells using Gradient Descent (GD) and Stochastic Gradient Descent (SGD) methods. Results show that the SGD method outperforms the GD method in terms of computational complexity. Both methods give similar performance for error in distance estimation with an accuracy of 0.89 m. The average localization error with prior knowledge of the location of MBS and femtocell is 2.35 m. Additionally, we derive the textbook derivation of Cramer Rao Lower Bound (CRLB) to analyze the performance of the estimator. The method successfully localizes smartphones in a dynamic environment, thereby making it suitable for 5G localization.

Indian Language Identification Using Cochleagram Based Texture Descriptors and ANN Classifier

Abstract: For a human translator, to identify many languages and to translate them accurately is a difficult task. The topic of identification of a spoken languages has become an active research field in the community of speech processing. In this paper, a two-stage identification approach for six Indian languages viz: Marathi, Hindi, Telugu, Malayalam, Kannada & Manipuri is proposed. In the first stage, the approach converts the input speech signals into their corresponding Cochleagram visual representations followed by their feature extraction using four different texture descriptors - Local Phase Quantization (LPQ), Completed Local Binary Pattern (CLBP), Binarized Statistical Image Features (BSIF) & Gray-Level Co-occurrence Matrix (GLCM). In second stage, the system identifies the spoken language using artificial neural network classifier. Combining different textural descriptors resulted in highest accuracy of 95.36% evaluated using Indic TTS database.

Feature Normalization for Improving the Performance of Sleep Apnea Detection System

Abstract: Sleep apnea is a common sleep disorder characterized by intermittent cessation of breath during sleep. Diagnosis of this disorder requires, prolonged and expensive sleep study test. The unavailability of such high-end diagnosis setup in rural areas, makes such disorders undiagonised. This paper introduces a low cost system which can detect sleep apnea by a rural health worker independent of sleep state, using electrocardiography (ECG) and respiratory effort signal (RES). The baseline-system uses statistical features derived from heart rate variability (HRV) and respiratory rate variability (RRV) data, with SVM used as a backend classifier. The feature vectors extracted from ECG and RES, carries patient and stage specific variations that does not contain information about apnea condition. Any effort to minimize these variations on the feature vectors can improve the performance. We explore two approaches to minimize these variations in the input features to improve the system performance. In the first approach we used nuisance attribute projection (NAP) in which we consider these variations as nuisance, and removed the components that are adversely effecting the performance of the classifier. Individual systems that are patient and stage independent were developed up on performing NAP algorithm and we got 81.25% sensitivity, 68.75% specificity and an overall accuracy of 75% absolute. Further using covariance normalization (CVN) we obtained an improvement of 16% absolute in the overall accuracy compared to the baseline-system. We further combined the NAP and CVN, and did not find any encouraging results.

Lateralization of Brain During EEG Based Covert Speech Classification

Abstract: Since time immemorial, researchers have attempted to decode the working of the human brain and tried to replicate its functionality. Brain-computer interfaces based on electroencephalogram (EEG) signals provide accessible and non-invasive means to capture the neuronal activity within the human brain. Towards analyzing the covert (imagined) speech, the brain activity during its production is captured through EEG signals. The ability to accurately interpret covert speech from EEG signals would be advantageous for communication by the patients suffering from speech disabilities or locked-in syndrome. On referring to the cognitive literature, we come across the Wernicke-Geschwind model that localizes the areas in human brain which are active during speech production and perception. In this paper, we present some classification results which are in support of the lateralization of the brain while analyzing the covert speech. For this study, a recently reported database which consists of EEG signals corresponding to pronounced and covert speech for six Spanish commands is used. A comparative study has been done between channels corresponding to the two hemispheres of the brain. The results indicate that there exists a laterality in brain functions and the dominant (left) hemisphere remains more active as compared to the right hemisphere during the covert speech similar to case of normal speech. This fact can be verified based on classification tasks performed on covert speech.

Ferroelectric Dual Material Gate All Around TFET Architecture for Enhanced Electrical Performance

Abstract: This work examines the device behavior of ferroelectric dual material gate all around tunnel field effect transistor (FE-DMGAA-TFET). The ferroelectric material manifests the negative capacitance that increases the drain current by developing an internal positive feedback on the gate terminal. The enhanced electrical characteristics of FE-DMGAA-TFETs is observed in contrast to DMGAA-TFETs. The device performance is examined by varying the thickness of ferroelectric layer and ferroelectric material. It is observed that on increasing the thickness of ferroelectric layer, improved device characteristics are acquired. Moreover, the results depict that FE-DMGAA-TFETs is a better contender for future generation and low power applications.

Design and Analysis of Logic Encryption Based 128-Bit AES Algorithm: A Case Study

Abstract: In today’s world, globalization of semiconductor industry leads to piracy and overbuilding of integrated circuits (ICs) and making it accessible for untrustworthy elements to insert hardware trojans. The solution to this problem is logic obfuscation which is used to encrypt the design by inserting additional gates so that valid output is produced only when the valid key is applied to these additional gates. This paper discusses the design and analysis of logic obfuscation based 128-bit AES algorithm. In this paper logic obfuscation based 128-bit AES algorithm is designed and simulated using Xilinx Vivado 2016.2. Results show that with logic obfuscation 128 bit AES algorithm offer the higher level of security and implementation flexibility with the small area overhead of 0.88%. The robustness of the proposed algorithm is also analyzed in the form of throughput and efficiency. The authors also presented the power consumption and analyzed the power overhead of 0.015W of the proposed algorithm using basys-3 FPGA. For the security analysis, hamming distance and avalanche effect have been used as evaluation metrics. Specifically, the proposed technique maximizes the uncertainty for an attacker as this technique manage to achieve 50% hamming distance and 40% avalanche effect between the correct and incorrect outputs when an invalid key is applied.

Human Activity Recognition with Wearable Biomedical Sensors in Cyber Physical Systems

Abstract: Human Activity Recognition has a wide range of applications such as remote patient monitoring, assisting disables and rehabilitation. This paper investigates the use of wearable bio-medical sensors to recognize human activities using supervised learning algorithms in cyber physical systems. We use five bio-medical sensors such as ECG, EMG, Respiration, Force sensitive resistor and a Tri-axial Accelerometer to collect the raw data. All the sensor data is collected in the real-world environment with three human subjects. The received raw data is preprocessed to extract the time domain features. The feature information is used for the training and testing the classifiers. Three classifiers k nearest neighbour (kNN), SVM using the linear kernel and SVM using Gaussian kernel are used for training and testing phases. The kNN classifier provides good accuracy of 99.86 %.

A Probabilistic Load Flow with Uncertain Load Using Point Estimate Method

Abstract: This paper presents the effect of shunt capacitors in improving the bus voltages and power losses in distribution system under stochastic load scenarios. Here, normal density function is used to model the uncertainty in load demand and 2 Point Estimate Method (PEM) and 3 PEM are used to determine the real power losses and nodal voltages. The probability density function and cumulative distribution function of the power losses and weak bus voltages are obtained by using Cornish-Fisher expansion series. The proposed method is tested on IEEE-34 bus radial distribution system. Two case studies are performed to analyze the positive impact of shunt capacitors in distribution systems under stochastic scenario. The results are compared with those obtained by Monte Carlo Simulation approach.

Design of Low Dropout Voltage Regulator for Battery Operated Devices

Abstract: A Low Dropout Voltage Regulator (LDO) provides a stable supply to an integrated circuit with a minimal dropout voltage across the pass element. This makes this circuitry very useful in various battery operated applications, which aims at portability and low power. It uses a Bandgap Reference (BGR) circuit to provide a temperature and supply insensitive reference voltage to the LDO,s error amplifier (EA). A basic configuration of a LDO along with a cascode BGR circuit with temperature coefficient of 183 ppm/°C to provide a temperature and supply independent constant voltage was designed and implemented in UMC 90 nm CMOS technology to provide a regulated supply of 1.2V. The LDO can provide upto 50 mA of current under full load and consumes only 10 μA quiescent current under no load, thus making it suitable for battery operated applications. Monte Carlo simulations show the design is immune to process and mismatch variations.

Geometrical and Structural Features for Forensics in Handwritten Bank Cheques

Abstract: In handwritten documents like bank cheques, adding new words can lead to financial loss. Moreover, identification of such words becomes more challenging when number of added words are very less. Since these documents are processed in digital form on daily basis, problem becomes even more worse to locate and to identify such changes in digital copy of these documents. In this regard, we propose a classification based approach that aims to identify whether pair of words are written by the same person or not. For this purpose, we use geometrical and structural features like stroke width, direction, inter character space, and so on. We use five classifiers, namely kNN, Decision tree, Radial basis-SVM, Multi-layer Perceptron (MLP), and Random Forest to train the model. The numerical results reveal that Random Forest classifier outperforms the other classifiers. The proposed scheme performs efficiently on IAM and IDRBT data sets using Random Forest classifier with average accuracy of 83.92% and 70.93% respectively.

Analysis of Global Bifurcation and Chaotic Oscillation in Distributed Generation Integrated Novel Renewable Energy System

Abstract: In this paper, a distributed generation (DG) integrated novel renewable energy system is proposed and analysis of bifurcation, chaotic behavior is observed. The integration of DG is the main property of renewable energy system, so in this model wind turbine generator (WTG) is integrated as a DG. Furthermore, dynamical load and Flexible AC Transmission System (FACTS) devices are also integrated. Renewable energy system is very complex, nonlinear dynamical system and it may exhibit chaos phenomenon in case of small, large disturbances or system parameter variations. The study of bifurcation and chaos behavior is essential in the proposed renewable energy system because it may lead to very serious results like voltage collapse, angle instability sometime blackout. The observation of bifurcation and chaotic oscillations are analyzed for wide range of parameter variations with the help of Lyapunov exponent, Lyapunov spectrum and bifurcation diagram. Simulation are done in MATLAB environment and simulation results reveals that the objective is achieved successfully.

MCSA: A Multi-constraint Scheduling Algorithm for Real-time Task in Virtualized Cloud

Abstract: Green cloud computing is the latest research trend where various approaches are introduced to minimize the energy consumptions and carbon footprint produced by computers. Further, the pay-per-use pricing model used in the cloud system helps to achieve economy of scale. Noticeably, many real-time applications that demand both temporal and functional correctness are moving to the cloud. It becomes a challenging task for a cloud service provider to ensure real-time response while minimizing computation energy and execution cost. In this regard, task scheduling plays a key role in achieving a performance improvement of the system with several constraints. In this paper, we proposed a multi-constraint scheduling algorithm, namely MCSA for the real-time task in the virtualized cloud environment. First, we assign a score value to a VM based on computation energy and execution cost. Then, MCSA use this scoring value to choose the appropriate VM for a task. We analyzed the proposed MCSA algorithm through extensive simulations and experiments. We consider Guarantee Ratio, Average Execution Cost, and Average Energy Consumption under various scenarios to show the effectiveness of MCSA over some existing schemes.

System Independent Machine Fault Diagnosis Using Convolutional Neural Networks

Abstract: A unified fault-modeling approach for machines with varying capacity, rating, size is of interest in automating production facilities. Building a unified fault-model is challenging due the presence of system specific attributes in the features derived. We explore how the effect of system specific attributes from the feature vectors can be minimized to improve the performance of system independent fault models(SIFM). In this work, we used three-phase synchronous generators with 3kVA and 5kVA ratings, for detecting the stator inter-turn short circuit faults. Our baseline-system uses statistical features derived from the current signal, with a support vector machine(SVM) used as a backend-classifier.In the first approach, we consider the system specific attributes as a nuisance and remove it using nuisance attribute projection (NAP) algorithm. We obtained a performance improvement of 19.63%, 8.31% and 11.80% classification accuracy for R,Y, B phases respectively, over the baseline-system.When we use a SVM backend-classifier, it is required that the features match the kernels used with the SVM. This often limits the performance of the classifier. Next, we explored convolutional neural networks(CNN) with the raw current signal as its input. Max-pooling is a technique used to minimize the size of feature map, simultaneously it construct a condensed feature map by selecting translation-invariant features. Thus, CNNs have the ability to absorb the irrelevant system specific variations, and improve the SIFM performance. It is observed that the CNN based SIFM outperforms NAP-SVM model by 12.56%, 10.37% and 20.15%, for R, Y, B phase fault models respectively.

Multimedia Forensic Detection Using Enriched Statistical Analysis

Abstract: Now- a- days, Internet has been used widely, law enforcement authorities have concerns regarding the trafficking of illicit materials like web images and videos, providing privacy and confidentiality to those materials is a major concern. Image data transaction over internet has become a major threat as those images contain hidden payload. People have desire to keep certain sensitive communications secret for thousands of years. In or new age of digital media and internet communications; this need often seems even more pressing. There is a need to identify the Stego data. The art of detecting the messages using steganography referred as Steganalysis. Developing such an application that identifies whether the selected image is a Stego image or not is highly recommended which can scan, detect and notify the user about the presence of payload. This paper is to identify the Stego image, non-Stego images, Stego videos and non-Stego video that can select the hidden payload. The main strength of this work is analysis and testing the proposed approach which has used to detect the hidden payload.

Gravitational Search Algorithm Optimized State Feedback Load Frequency Controller

Abstract: To maintain reliable and stable operation in a power system, maintaining the value of frequency variation within a desired limit is important. In this paper, a state feedback controller is designed for a load frequency control (LFC) system. This process of defining each state will help the operator to understand the cumulative behavior of different states on the system performance in presence of external disturbances. A generalized range for stability is derived depending on the physical parameters of the system. Instead of manually selecting the controller gain values from stable range, an optimization algorithm is used. In this paper, gravitational search algorithm (GSA) is used as an optimization algorithm. GSA minimizes the effect of external disturbance on system frequency and gives a set of controller values satisfying system stability criterion. A set of results is presented to show the improved closed-loop performance of the system. A comparative study with the previous work of the authors and with a few existing works is also presented to highlight the advantages of the proposed method.

Throughput and Energy Efficiency of Cooperative Cognitive Radio Network over Erroneous Generalized Fading Channel

Abstract: In the present paper, we propose an energy detection based cooperative cognitive radio network (CCRN) with hard-decision combining fusion. All cognitive radio (CR) nodes sense the primary user (PU) using erroneous sensing channels and report their sensed data to the fusion center (FC) using erroneous reporting channels in the form of local one-bit binary-decision. At FC, the final decision about the PU (presence or absence) is obtained using n-out-of-N fusion rule. Specifically, we derive an expression for detection probability under noise and generalized κ − μ fading channel and the mathematical model for energy efficiency of CCRN is developed. Further, for different values of the CCRN parameters, the performance subject to various decision fusion rules is studied. The analytical performance characteristics are investigated mainly via total error probability (TEP) and energy efficiency over noise and generalized κ − μ faded channels. Finally, optimized values of the required network parameters are determined and optimal subrule which maximizes the energy efficiency is also identified.

A Transfer Learning Approach for Diabetic Retinopathy Classification Using Deep Convolutional Neural Networks

Abstract: Diabetic Retinopathy is a disease in which the retina is damaged due to diabetes mellitus. It is a leading cause for blindness today. Detection and quantification of such mellitus from retinal images is tedious and requires expertise. In this paper, an automatic identification of severity of Diabetic Retinopathy using Convolutional Neural Networks (CNNs) with a transfer learning approach has been proposed to aid the diagnostic process. A comparison of different CNN architectures such as ResNet, Inception-ResNet-v2 etc. is done using the quadratic weighted kappa metric. The qualitative and quantitative evaluation of the proposed approach is carried out on the Diabetic Retinopathy detection dataset from Kaggle. From the results, we observe that the proposed model achieves a kappa score of 0.76.

Optimal Placement and Sizing of Harmonic and Reactive Compensators in Interconnected Systems

Abstract: Integration of distribution generators such as solar photovoltaic, wind, geothermal etc. is necessary to meet the huge and dynamic power demand in interconnected systems. High penetration of these distributed generation resources causes varying reactive power demand, grid voltage instability, transmission congestion, harmonics, large power losses, reduction in source power factor and other imbalances in the net-works. These issues will be propagated throughout the system and affects quality of power supplied to all other loads in the system. Hence installation of effective compensators is necessary. This paper presents bacterial foraging based optimal allocation and selection of compensators in an interconnected system. Static synchronous compensator, thyristor switched series compensator, passive and active filters are used as compensators in the system. The algorithm is tested in MATLAB/SIMULINK with IEEE – 5 bus system and results are compared with other combinations of compensating equipments.

Design and Development of Opto-Resistive Type Quasi-digital Sensor and Instrument to In-situ Monitoring of the Quality of Lubricant Oil

Abstract: In large rotating machinery, lubricating oils are employed to reduce the wear, minimize the friction between gears and bearings, remove heat, takeaway wear particles and to protect against corrosion. Lubricant oil slowly degrades over a period of time due to oxidation and contamination. Consequently looses its defensive properties and effects the machinery performance. This may leads to catastrophic failure, unplanned shutdown and capital loss. Replacing the in-service oil with fresh oil before degradation can prevent the above consequences but changing the lubricant too early is uneconomical. Therefore, it is better to predict the remaining effective life of the oil under operating conditions. Monitoring the healthiness, predicting the remaining useful life of the lubricant oil and warning the operator with an alarm can prevent catastrophic failure. It is observed that the oil color becomes darker with contamination, oxidation, and aging. Thus, the color of the lubricant is considered as an essential parameter for describing the quality of the oil. This paper describes the design and development of a lowcost, low power instrument and Opto-resistive quasi-digital sensor. It explains the capability of the instrument about in-situ monitoring of quality, predicting the effective life of lubricant by measuring the transmittance of the light through lubricant oil and alerting the operator by annunciation cum display.

Effect of Varying Temperature on Different PV Array Configurations Under Partial Shading Condition

Abstract: The aim of this paper is to study the effect of varying temperature on different photovoltaic (PV) array configurations: Series (S), Parallel (P), Series-Parallel (SP), Total-Cross-Tied (TCT), Bridge-Linked (BL), Honey-Comb (HC) and proposed hybrid configurations: Series ParallelTotal Cross Tied (SPTCT), Bridge Linked-Total Cross Tied (BLTCT), Honey Comb-Total Cross Tied (HCTCT) and Bridge Linked-Honey Comb (BLHC) under partial shading condition. The results for these PV array configurations, under partial shading condition, corresponding to a general shading pattern with bypass diodes by varying temperature from 0° C to 75° C generated in MATLAB are obtained. The analysis of effect of varying temperature on these PV array configurations has been accomplished by considering PV arrays of dimensions 6 × 4 and 8 × 4, made of KD260GX-LFB2 PV modules. The P — V characteristics generated in MATLAB for different PV array configurations, under partial shading condition, corresponding to a general shading pattern with bypass diodes by varying temperature are validated against the experimental data of PV arrays of dimension (8 × 4), made of KD260GX- LFB2 PV modules.

Performance Investigation of Interference Alignment Techniques for Underlay MIMO Cognitive Radio Networks

Abstract: Underlay cognitive radio network allows secondary users to use the primary user spectrum as long as the interference caused by the secondary user is below a certain threshold. Interference Alignment (IA) algorithms were proposed to mitigate the interference thereby enabling simultaneous transmissions from multiple users over same frequency band. This paper presents the comparison of the performances of four different IA algorithms: Minimum leakage Interference Alignment, Maximum SINR Interference alignment, Partial Interference Alignment considering secondary users and Partial Interference Alignment not considering secondary users, in underlay Cognitive radio scenario.

Comparative Evaluation of PMU Algorithms for Power Quality Monitoring in Distribution

Abstract: By increasing the utilization of the grid integrated renewable power plants and use of nonlinear characteristic appliances, the quality of power is deteriorated. The increase of complexity in the power system affects power quality appreciably. To monitor the state of the distribution system, India is planning to install 1700 phasor measurement units (PMU) in the near future. Up to what degree the PMU will benefit the power quality monitoring is not known. In this paper, the investigation of the PMU algorithm and comparison with the classical algorithm are made to detect the dips and swells in the voltage profile. Till now PMUs are not in use to detect the power quality of the supply. The input data used in this paper has not been taken from the real system and all the algorithms run in this scenario to a simplified comparison between different algorithms. The Presented work will lighten the possibilities and the limitations of the PMU algorithm for swells and dip detection in the power supply.