National Institute of Technology, Meghalaya

About: National Institute of Technology, Meghalaya is a education organization based out in Shillong, India. It is known for research contribution in the topics: Control theory & Computer science. The organization has 503 authors who have published 1062 publications receiving 6818 citations. The organization is also known as: NIT Meghalaya & NITM.

Statistical analysis of oil insulation breakdown voltage

Abstract: Liquid or oil insulation are widely used in electrical power apparatus. The breakdown characteristic of oil insulation is statistically analysed. In order to obtain adequate information, suitable statistical distribution function needs to be identified to map with breakdown voltage data set. In this work, breakdown test of oil insulation is carried out with certain ramp rate i.e. rate of rise of voltage. The experimentally obtained breakdown voltage data was mapped with normal and weibull distribution. The statistical distribution of data is adopted to obtain information about the breakdown characteristics of oil insulation. The normal distribution is used to estimate mean value, standard deviation, whereas scale parameter, shape parameter are estimated using weibull distribution. Skewness factor is estimated for normal distribution and is seen to be of non-zero value. This non-zero skewness represents asymmetric distribution of experimental data. Weibull distribution being asymmetric in nature fits well with the experimental data. The weibull distribution provides significant information about the breakdown mechanism as well as characteristic breakdown voltage magnitude.

Identification of Significant Frames from Colonoscopy Video: An Approach towardsEarly Detection of Colorectal Cancer

Abstract: Colonic polyp detection during colonoscopy is an essential and crucial task towards the improvement of automatic detection of colon cancer. Advancement of technology in the field of medical image analysis helps to identify anomalies in its early phase and helps in the therapy planning process. The work presented here is a methodology focused on deep learning techniques that give an edge to the identification of the existence of polyps. This proposed approach could serve as an assistive tool during the process of colonoscopy.

Modelling and Hardware-in-Loop validation of a modified controller for intermittent operation of a standalone low voltage DC fuel cell system

Abstract: Integration of renewable energy into conventional power system is increasing day by day which is motivating researchers to explore new forms of renewable energy and use them efficiently to fulfil generation demand gap. This paper deals with the design and operation of a standalone DC microgrid using fuel cell stack as a source of energy, coupled to a DC resistive load through a boost converter whose output voltage is controlled by a dual loop PI controller. The proposed system is designed to output a stable DC voltage of 48 V irrespective of the load variations and slow starting dynamics of fuel cell stack. Modifications made in controller configuration to facilitate the intermittent operation of the fuel cell is explained in detail. The modified controller configuration improves the reliability of the fuel cell system irrespective of the intermittent nature of the load. Time-domain analysis of the above said scenario is carried out using MATLAB/SIMULINK and is validated by a Hardware-in-Loop analysis in Virtex-6 FPGA using Xilinx system generator.

Comparative Analysis of Vector Controlled PMSM Drive with Particle Swarm Optimization and Ant Colony Optimization Technique

Abstract: This paper presents performance analysis of two optimization techniques to control speed of the Permanent Magnet Synchronous Motor (PMSM). Here, Particle Swarm Optimization (PSO) and Ant Colony Optimization (ACO) method are used to determine the integral gain and proportional gain of PI (proportional integral) controller. For the analysis, simulation is carried out in MATLAB/SIMULINK for three different cases such as constant speed with no-load condition, variable speed at no-load condition and constant speed with loading condition. Different parameters such as torque ripple, peak time, settling time, rise time and maximum overshoot are analyzed in this paper for both the optimization technique. From the result it can be observed that ACO tuned PI controller in PMSM drive give better performance comparing to PSO tuned PI controller under different condition.