SAMRIDDHI - A JOURNAL OF PHYSICAL SCIENCES, ENGINEERING & TECHNOLOGY

Design of Compact Bow-Tie Slotted Antenna for Satellite Communication

Mamatha A G, Pradeep M Hadalgi — Mon, 26 Dec 2022 00:00:00 +0100

A printed antenna using slot loading technique is adapted in the design to achieve multiband. A rectangular microstrip (printed) antenna with a Bow-Tie shape slot embedded on the radiating patch using a glass epoxy substrate with dimensions 3.34 x2.47 x 0.16 cms. The antenna is simulated using 3D electromagnetic simulator software HFSS and fabricated using photolithography. Vector Analyzer is used to measure the experimental results. The fabricated antenna has return losses of -20.00 dB, -25.09 dB, and -26.45 dB at three independent frequencies 4.62 GHz, 8.14 GHz, and 12.41 GHz. The antenna operates at a frequency range of 4.62 GHz to 12.41 GHz, which supports satellite (space to earth) 5.441 and mobile communication 5.463 applications.

Failure Analysis and Field Repair of Urea reactor

Krishnan Sivaraman, Ravinder Singh, Abhishek Singh — Mon, 26 Dec 2022 00:00:00 +0100

The outer shell of a Multi-layered Urea reactor ruptured while the plant was in operation. The plant madean emergency shutdown, samples were collected, and an investigation carried out to establish the root cause of the failure. The paper explains the detailed analysis performed, corrective action implemented, and a health assessment for the equipment’s safe working. The paper also includes references to similar failures in old multi-layered urea reactors subsequently and the mitigation plans to avoid such failures

Non Newtonian Model for Two Phase Blood Flow in Hepatic Arterioles in Case of Dengue using Herschel –Bulkley Law

Bhaskar Pandey, V Upadhyay, A K Agrawal, P N Pandey — Mon, 26 Dec 2022 00:00:00 +0100

Herschel-Bulkley Non-Newtonian model has been studied taking blood flow in arterioles during dengue which have two phases. First one is a blood plasma and other one is RBC. Equation of continuity and equation of motion for two phase blood flow have been derived in tensorial form with respect to blood flow in hepatic circulatory sub-system. Model has been applied in biomechanical setup. A clinical data is taken in the case of dengue. Numerical methods are followed for finding the parameter. Relationship between hematocrit and blood pressure drop is also showed by graph.

Data Acquisition and Remote Monitoring of Geological Parameters for Landslide Prediction using WSN

D Srinivasa Rao, Mohammed M Ahmed, C Vinay Kumar, Sake P Kumar — Mon, 26 Dec 2022 00:00:00 +0100

Aiming at the problems of a large amount of redundant information, low data accuracy and difficulty in system monitoring when wireless sensor network nodes collect data. An efficient technique of data collection and remote monitoring system based on wireless sensor network is proposed. The system uses WSN nodes to real-time collection of external environment parameters. A software design for Wireless Sensor Network formation, data fusion, data transmission and transferring data to the host computer, and use of WIFI/GSM as a means of transmission to build a remote monitoring platform. The test results show that the system can realize remote real-time monitoring of the external environment, geophysical parameters and the use of data fusion algorithms increases the transmission rate and data accuracy between nodes.

Relative Entropy-based Energy Efficient Multi-hop Forwarding Routing Protocol for Wireless Sensor Network

Akant K Raghuwanshi, Virendra S Chaudhary — Mon, 26 Dec 2022 00:00:00 +0100

Adaptive Clustering Hierarchical Protocol (LEACH), followed by cluster head election. Conversely, it is not thought that data redundancy is sufficient. This study suggests a modified Leach approach incorporating relative information entropy (Leach-CIE). The relative entropy value of the data’s probability distribution in the node’s two subsequent periods is calculated in conjunction with the close entropy model. The information is then compared with a threshold value to determine the degree of redundancy. The network accomplishes reducing energy consumption by refusing to transmit redundancy data. Instead, a multi-hop routing (MFRCRE) considers its residual energy, the energy consumption of each node in the network during the data transmission phase, by optimizing node forwarding conditions and considering factors such as communication distance and energy consumption. It is proposed to balance the rate—account ratio. The results of the experiments demonstrate that this approach can successfully reduce and balance the energy consumption of network nodes to increase the life of the network.

Resource Allocation in Massive Internet of Things-Edge Network with Optimal Path Planning and Scheduling

Ajay K Barapatre, Ritesh Sadiwala — Mon, 26 Dec 2022 00:00:00 +0100

Although an Ultra Dense deployment is required for 5G services, it will be nearly impossible to achieve 60 service coverage with the dense deployments due to the even shorter transmission range. In light of 5G’s impressive technological advance, 5G and data centers are now capable of handling an increasing number of real-time and complicated computational tasks from Internet of Things (IoT) systems. This paper proposes an optimal mobile resource-sharing approach to confront this underlying limitation of 5G. In contrast to conventional algorithms, the designed optimal path planning and scheduling for mobile edge server (OPPSMES) is proposed that have the advantage of a synchronization among request being received and achieved lower delay and resource demand for as computing this allowed for the parallel processing of task and server in mobile condition. The OPPSMES includes two steps, i.e., path planning and optimal task scheduling, to improve efficiency. According to simulation outcomes, there is a significant increase in resource utilization and a decrease in average response time.

The Hash base Apriori Technique for Association Rule Mining and Data Sanitization

Nitin P Jagtap, Krishankant P Adhiya — Mon, 26 Dec 2022 00:00:00 +0100

Under different circumstances, private information is exposed, and it must be sanitised before even being shared to address privacy issues. Data mining techniques can collect large amounts of data in a short amount of time. The information gathered by the powerful machine learning techniques may identify the most sensitive content, which pertains to an individual or organization. The degree of sensitivity of data belonging to a business or an agency might vary. Only approved individuals and organizations have access to this information. As a result, using access limitations to confirm the security of complicated data is not a complete operation. It can impact the utility of a data mining solution, and the user may be able to re-identify sensitive data. To introduce instruments to find a mechanism for the security of confidential information. Finding ways to secure confidential data by developing data mining tools and procedures that can be applied to databases, even though this diminishes the data mining results’ trust worthiness. In this article, we proposed a data sanitization strategy that uses a frequent itemset classification approach with a modified apriori algorithm. The problem is to maintain intelligence information for vital arrangements while simultaneously preventing the numerous exposures of company rule mining. Data sanitization strategy is used to thoroughly investigate numerous sequential pattern algorithms for ensuring the privacy of large amounts of data. Our research shows that our approach is efficient, scalable, and provides meaningful correction compared to other methods used in existing systems

Enhancement of Power Quality with Increased Levels of Multi-level Inverters in Smart Grid Applications

Abhimanyu Kumar, Sanjay Jain — Mon, 26 Dec 2022 00:00:00 +0100

Research focuses on renewable energy-based smart grids to meet appropriate energy demands in a world where climate change and power management are becoming more significant. The smart grid is a modernized autonomous power network that can efficiently transmit electricity, save money and resources, and improve the stability of the local grid. As a result, a multi-level inverter (MLI) that is connected to the smart grid is provided in this work. To meet the requirements of the smart grid network, the inverter is controlled using a model predictive control algorithm with increased levels with the main objectives of controlling the injected power generated by the renewable source, improving the quality of the current waveform, lowering THD, and eliminating the shift phase between the current and the grid voltage

Carbon Capture from Coal Power Plants and Algae Farming with Technical-Economic Analysis

Amit Sharma, Nitin Dubey — Mon, 26 Dec 2022 00:00:00 +0100

55% of global warming. According to research, the quantity of CO2 in the atmosphere corresponds well with the growth in the average global temperature of the planet's surface. The main aim of this work is to investigate different emerging Technologies for the carbon capture from power plants and its atmosphere. After capturing of CO2, it is used for the injecting in raceway ponds for micro algae cultivation. Mathematical and experimental analysis for the carbon capture and Sequestration from the coal power plant have been performed in the RKDF University and proposed the microalgae production using captured CO2. For that total 24 experiments were performed to get the efficiency of the CCS plant. The result shows that Input of 250 kg/ hr of flue gasses has been used to capture the CO2. Sodium Carbonate solution of 10% will be sprinkled on top of the packing in tower-A to increase the contact area. The gases are now subject to the counter-current of MEA solution of 30% contained with 1 kg of CO2. MEA-CO2 mixture is heated up to 155°C with the help of a Scheffler dish collector. CO2 gas released from the MEA-CO2 mixture when heated above 130°, the maximum efficiency of the CCS plant of 87.58 % on 04 Jan 2021 at 1200 Noon.

CBR Characteristics and Swelling Behaviour of Black Cotton Soil Mixing with Fly Ash and Silica Fume

Beerendra Kumar, Ashish Gupta, Adarsh Kasaudhan — Mon, 26 Dec 2022 00:00:00 +0100

Today’s rapid growing world produces waste that is very difficult to handle eco-friendly. Soil contamination is also one of the major geo-environmental issues by pouring industrial wastes directly over soil without treating it like caustic soda, fly ash, etc. Fly ash reaches ground water through percolation and pollutes the ground water. The main reason behind the soil contamination is due to rapid urbanization. This innovative and non-traditional research on waste utilization is gaining importance nowadays. Geotechnical engineers are adopting soil properties remolding and monitoring techniques by using waste material like fly ash Slag, Silica fume etc. This research aims to find the suitability of by-product silica fume and fly ash for Black cotton soil. In this research number of laboratory test are performed on black cotton soil using silica fume & fly ash as admixture in varying percentage. The outcome of the research shows the enhancement in properties of black cotton soil also significant decrease in free swelling index is observed. This expansive soil has very high swelling and shrinkage characteristics due to the clay mineral montemorillonite. It is found in abundant quantity in southern region of India and due to its poor characteristic, it is also termed as poor soil. Building any structure, road, bridge, embankment etc is so risky, and the solution is to replace this poor soil with good natural soil. Hence in this thesis work, emphasis is given on the use of black cotton soil by mixing stabilizers like silica fume and fly ash. Both increase the maximum dry density of black soil, and also CBR %value is increased. On adding silica fume, it also possesses a very low free swell value; hence, using both additives simultaneously will be a better solution to the problem of black cotton soil.

Design of Multi Input Converter Topology for Distinct Energy Sources

Chirag Gupta, Vikas K Aharwal — Mon, 26 Dec 2022 00:00:00 +0100

Two input energy sources are introduced in the suggested Multi input DC-DC converter topologies. The converters may send power to the load from both the input and output energy sources at the same time. The capacity to execute the buck, boost, and buck-boost modes of operation using the same structure, as well as the ability to transmit power to the load even if any one of the input energy sources fails, are the key advantages of the upgraded converter over the basic design. As a result, the MATLAB/Simulink platform was used to do a full software simulation of the enhanced converter. The state topology of the converter is illustrated independently for the buck, boost, and buck boost operations, indicating the integration of two renewable energy sources. In comparison to current converter topologies that have been published in the literature, the suggested converters offer various advantages such as a lower component count, a compact construction, and efficient energy usage

An Approach to Prevent Neighborhood Attack over Social Media

Jitendra Patel, Ravi K S Pippal — Mon, 26 Dec 2022 00:00:00 +0100

Social media sites contain the personal information of the users, which entice the attackers The attacker uses several types of attacks on the social networking site to obtain sensitive information from the users. As different types of passive and active attacks are carried out on social media sites, user privacy may be compromised; to avoid this, the network operator releases data anonymously. Data from social media users is gathered and stored by social media operators for distribution to various third-party consumers. Because the fetched data frequently contains sensitive information, the network operator makes the entire graph available in anonymized and sanitized forms. It does not, however, provide a complete guarantee of user privacy. This research provided a way for anonymizing social network graphs using a neighborhood adjacency matrix-based anonymization process. This anonymization procedure could be utilized to defend against the social network graph's neighborhood attack. By adding fake edges to the social network graph, the suggested anonymization procedure increases the number of isomorphic neighborhood networks. As a result, a user's unique neighborhood network cannot be used to re-identify them in a social network graph.

Preparation and Dielectric Investigations of TiO2 mixed ZnO-TeO2-B2O3 Glasses

Sopinti Ramesh, Gunnam Nagarjuna, Bejjipurapu Chandrasekhar, B V Raghavaiah — Mon, 26 Dec 2022 00:00:00 +0100

Glasses doped with 0.3 to 1.2 mol% of TiO2 doped ZnO-Sb2O3-B2O3 are synthesized. Samples were tested for their nature by XRD and morphology by SEM pictures. Dielectric constant, loss and a.c. conductivity was measured from temperature 30 to 200oC and frequency range 1 kHz to 100 kHz. The three dielectric parameters were found to rise with the quantity of titanium oxide. The trivalent titanium ions were found to be more in the sample ZTBT12 and have shown the highest dielectric constant and loss, however activation energy is small.

Feature Reduction in Classification Tasks using Bio-inspired Optimization Algorithms

Rachna Kulhare, S Veenadhari — Mon, 26 Dec 2022 00:00:00 +0100

In big data, there is a major difficulty that requires data mining to be conducted with elevated data in big technology, which would be gaining a lot of traction nowadays. When it comes to Big Data, feature selection approaches are seen to be a game changer since they can assist minimize the complexity of data, making it simpler to study and translate it into meaningful information. To enhance classification performance, feature selection removes unnecessary and redundant characteristics from the dataset. In this paper, Grey Wolf Approaches based on Quantum leaping neighbor memeplexes termed as QLGWONM is proposed. The result shows that when compared to the some bio-inspired algorithms such as PSO, GWO, ABA, CSA models, the suggested model performed well in terms of accuracy and have accuracy of 100% for brain tumor, CNS, Lung dataset and 97.1% for Ionosphere dataset and 99% for NSL-KDD.

A Secure Communication Protocol for Unmanned Aerial Vehicles using IoT Protocols

Jayant kumar Dorave, Ritesh Sadiwala — Mon, 26 Dec 2022 00:00:00 +0100

IOT_MQTT is an open-source communication system for Unmanned Aerial Vehicles that is lightweight and extensively utilized. It is supported by a number of UAVs and autopilot systems, and it allows bidirectional communication. The findings demonstrate that our approach can encrypt IOT packets while maintaining performance and efficiency. The findings are confirmed in terms of transfer speed, performance, and efficiency when compared to other solutions in the literature such as smart dust and benchmarked against the original IoT-MQTT protocol.

Lung Infection Detection using Progressive U-NET Architecture

Ashish Pandey, Sandeep Dubey — Tue, 28 Mar 2023 00:00:00 +0200

The fragmentation of medical images of tissue anomalies, imorgans, or the blood vascular system is critical for any computerized diagnostic system. Nevertheless, automated categorization in clinical visual assessment was a difficult problem as it necessitates in-depth information about the specific organ structure. This article presents UNET, an edge deep learning categorization technique for early recognition of COVID. This is also a problem for treatment and correct intervention, as previous techniques were inappropriate in this situation. Trying to keep this in consideration, non-invasive methods such as CT scans and X-rays were suggested to get characteristics of lungs

Developing a Hybrid model with Shades of Sentiment for Understanding Teenagers’ Academic Distraction Problems

Monali K Patil, Nandini Chaudhari, B V Pawar, Ram Bhavsar — Mon, 26 Dec 2022 00:00:00 +0100

In recent years, many techniques have come up in the field of sentiment analysis. In the field of the medical domain, sentiment analysis has been used for the areas like information retrieval, feedback analysis, dialogue conversation, and review analysis. Psychological analysis through the window of sentiment analysis is still unfolded area. We have designed and developed a hybrid computational model that maps teenagers’ sentiments to their behavioral patterns and academic distraction problems. We have also developed a sentiment shaded lexicon which defines the ontology for various shades of sentiment with the help of Bling Liu’s positive and negative word dictionary. We have used a semantic lexicon-based approach and a rule-based classifier in our hybrid computational model. Rules are applied to user-input text, which teenagers write. We have extracted sentiments expressed in the user input text and we have also achieved to identify academic distraction problems of teenagers. We have computed the performance metrics of the model on 155 samples, randomly collected from teenagers (age group 13-19 years). We have computed accuracy at two stages, first at the sentiment extraction stage and second at the final output level. The error analysis is also presented in this paper. After working on error analysis, our model has achieved an accuracy of 90% for sentiment extraction and 87% for derived academic distraction problems.

Deformation Analysis in Aluminium Magnesium Alloy Using Digital Image Correlation

Mon, 26 Dec 2022 00:00:00 +0100

Image processing is extensively used to examine surface deformations. Digital Image Correlation (DIC) detects two-dimensional subpixel displacements in two images to analyze deformations in geomechanical structures. In this study, the analysis is done on an aluminium-magnesium alloy. An optical microscope is used to capture the image, which is then processed and utilized to anticipate material deformation. To acquire findings of strain points on the material, the Ncorr application in the MATLAB environment is employed. . Strain graphs on the Exx, Exy, and Eyy axes, as well as U-displacement and V- displacement graphs, are generated to detect the deformation.

Experimental Investigation of Recycled Aggregate Concrete Including Metakoline

Arpit Sethia, Shashi R Kumar — Mon, 26 Dec 2022 00:00:00 +0100

As construction is increasing day by day in last decade, it is concerned for processing of waste material with the help of recent technology. The current study investigated the effect of metakoline (MK) on mechanical properties of RAC. The current study presents the experimental results of Recycled aggregate (RA) with different percentage 0, 20, 40, 60, 80, 100% in concrete replaced by natural coarse aggregate. Casting of cube, beam, cylinder mould has been done. In addition to this replacement of cement partially with MK also has been introduced. The flexural strength, compressive tensile strength and split tensile strength of M-30 and M-25 have been determined and also there design mix has been prepared. The results show good effect up to 60% aggregate replacement but on further increases the replacement ratio of aggregate, strength decreases. Therefore sustainable concrete may be produced to make environment safe and eco-friendly

Design of a Hybrid Bio-inspired Model for Improving Addressing Capabilities of IPv6 Networks

Reema Roychaudhary, Rekha Shahapurkar — Mon, 26 Dec 2022 00:00:00 +0100

Addressing in IPv6 networks is a complex task, that involves multiple decision criterion that include base addressing, subnet addressing, location & energy aware addressing, and temporal performance aware addressing constraints. To incorporate these constraints a wide variety of models are proposed by researchers, and most them utilize location-aware addressing and do not consider multimodal parameters. Schemes that consider these parameters are either highly complex, or cannot be scaled for heterogeneous network scenarios. To overcome these limitations, this text proposes design of a novel hybrid bioinspired model that assists in improving addressing capabilities of IPv6 networks. The proposed model bee colony optimization (BCO), genetic algorithm (GA), and particle swarm optimization (PSO) in order to improve addressing quality for different network types. Initially, GA is used to stochastically assign location-specific addresses to a simulated network, which is incrementally tuned by PSO via a cognitive & social learning process. The fine-tuned addresses are further optimized via integration of BCO model, which assists in integrating energy awareness. Final addresses are initially simulated with exhaustive communication test, and then deployed to real-time networks for optimized operations. Due to which, the assigned addresses are observed to be delay & energy efficient, thereby assisting in deploying them for real-time use cases. The proposed addressing model was tested under different scaled networks, and an energy efficiency of 8.3%, with delay reduction of 6.5% was achieved when compared with various state-of-the-art methods, which assists in deploying the model for multiple scaled network scenarios

An Integrals Involving the Multivariable H-Functions

Mithilesh K Mishra, Rajeev Shrivastava, Lakshmi N Mishra, SK Tiwari — Mon, 26 Dec 2022 00:00:00 +0100

The aim of this paper is to evaluate an infinite integral involving the product of multivariable H-functions along with Srivastava polynomial and M-series by means of finite difference operations E. As the generalized hypergeometric function and multivariable H-functions are of a very general nature, the integral, on specializing the parameters, leads to a generalization of many results some of which are known and other are believed to be new.

Formulation and Implementation of Energy Efficient Routing Algorithm for MANET along with Performance Analysis

Amandeep Makkar, Mukesh Kumar, Sunil Taneja — Mon, 26 Dec 2022 00:00:00 +0100

Mobile adhoc network (MANET) is composed of wireless mobile nodes/devices which are moving randomly in infrastructure-less adhoc environments. Topology among mobile nodes is dynamic in nature and keeps on changing regularly. The adhoc environment is open to all the genuine users as well as intruders/attackers. Moreover, the wireless connection between mobile nodes can break at any time on account of mobility of nodes. Also, the mobile notes depend on their battery power, which may go down because of exhaust of battery power. Hence, it is a very complex task to implement energy efficient routing over MANET. While handling energy efficient routing, the optimum utilization of batter power is required. In this research paper, efforts have been carried out to formulate an algorithm for energy efficient routing over MANET. Its performance has been studied through network simulator. The simulations have been carried out by writing self created network scenarios with 20, 60 and 100 mobile nodes. Moreover, analysis of the proposed energy efficient routing protocol has been done with AODV routing protocol by using various performance evaluation metrics viz. packet delivery ratio, normalized routing load, average end to end delay, packet loss, throughput, average energy consumed and average energy left. It was derived from the performance analysis that proposed protocol provides energy efficient routing over MANET and also gives better or almost same performance in comparison to AODV.

On Certain Triple Integral Relations Involving Elementary Functions

Mon, 26 Dec 2022 00:00:00 +0100

The aim of this paper is to establish three triple integral relations involving elementary functions. A number of triple integrals can be deduced by proper specialization of the unknown functions f and g occurring in these relations. For the sake of illustration, one of our integral relations is applied to evaluate a general triple integral involving Asgar, Gautam and Goyal multivariable A- function.
Mathematics Subject Classification: 33C45, 33C60, 26D20

Smart Wearable System for Women Security using Internet of Things (IoT)

Pravin Pardhi, Chetna D Dhote, Riteshwari A Pagote, Suwarna S Somkuwar — Mon, 26 Dec 2022 00:00:00 +0100

Women safety is a big issue in even today’s modern world driven by advanced technologies. This function demonstrates an efficient women safety identification system using a GPS modem. The proposed women protection device aims to provide complete protection to women in the present scenarios. This device provides a safe and peaceful environment for women with handbag security and android application. This work aims to integrate the Internet of things (IoT) based women’s safety device with the device’s hardware linked in the handbag, Android application and Bluetooth connectivity in smartphone. After pressing the device key the Android Application identifies the location of the victim’s place via GPS and sends a message containing this location URL to the registered contact and also call on the first registered contact to help them in a dangerous situation and also send the alert message to the registered mobile number. location tracking information through SMS helps find out the victim’s current location quickly and can be saved securely.

A Review on Performance Analysis of Deep Learning for Task Offloading in Edge Computing

S Almelu, S Veenadhari, Kamini Maheshwar — Mon, 26 Dec 2022 00:00:00 +0100

Edge computing is a type of distributed computing that was designed especially for Internet of Things (IoT) users to provide computational resources and data management nearby to users' devices. Introducing edge computing for IoT networks has reduced the bandwidth and latency issue while handling real-time applications. The major benefit of edge computing is that it reduces the communication overhead between IoT user and server. Integrating IoT in our daily life has attracted researchers towards its performance management such as complexity minimization, latency minimization, memory management, energy consumption minimization, etc. The paper focuses on deep reinforcement learning to minimize the computational complexity at IoT user end. The task offloading decision process is designed using Q-Learning that minimizes the system cost and curse of high dimensional data.

Ultrafast Spectroscopy of Semiconductors: A Review on Experimental Methods

Richarj Mondal — Tue, 25 Oct 2022 00:00:00 +0200

The advancement of picosecond and femtosecond pulsed lasers made optical spectroscopy a more powerful tool that has been used in the fields of physics, chemistry, and biology. This method is mainly used to study the dynamical properties of various substances such as metals, semiconductors, and superconductors. Ultrafast non-equilibrium carrier excitation of the semiconductor and subsequent relaxation processes occurring on different time scales have become an important area of semiconductor research. Numerous experimental approaches have been employed to explore these dynamic processes in both semiconductors and their nanostructured counterparts. In this article, we have discussed dynamic phenomena that occur in optically excited semiconductors on time scales from a few hundred femtoseconds to nanoseconds. Among the most widely employed experimental techniques is the ultrafast pump-probe technique, which facilitates the exploration of dynamic phenomena in semiconductors. This article provides a comprehensive explanation of the basic principles that form the basis of the ultrafast pump-probe technique. Additionally, it highlights significant experimental discoveries obtained through the investigation of semiconductor quantum wells and quantum dots.

ULMFiT Embedding(s) for Context and Extended Gloss Intersection for Marathi Word Sense Disambiguation

Sandip S Patil, R P Bhavsar, B V Pawar — Mon, 26 Dec 2022 00:00:00 +0100

Ambiguities in the word meanings makes all the natural language processing (NLP) tasks very difficult, word sense disambiguation (WSD) is used to resolve these ambiguities. Now a day’s NLP-based human assistive systems are in demand, in which machines are expected to resolve word sense ambiguities. Today, due to the availability of machine readable dictionaries knowledge-based WSD approaches have become popular; it explores semantic relations between the contextual features and possible glosses of the given ambiguous word. Inductive transfer learning-based language models have great potential to represent the different semantic features of the word, which can be used in various NLP tasks. Universal language model fine-tuning for text classification (ULMFiT) is a popular transfer learning model used to embed various semantic features in digitally resource scare and morphologically rich language like marathi. In this reported work, the ambiguous words from the Marathi input sentence is extracted and have obtained its possible synset and glosses from IndoWordNet, these glosses are then extended using hypernym and hyponym relations. We have obtained the word embedding of marathi context and extended glosses using ULMFiT model. For the test run, we have crafted the test-bed of 6000 marathi sentences of 280 moderately ambiguous words harvested from marathi websites, which caters for 1200 senses. The winner sense is declared based on the maximum intersection score between the pair of context and gloss embedding. We have obtained the average accuracy up to 57.10% for our dataset.